EVALUACION DE VIABILIDAD DE LA POBLACION Y DEL HABITAT DEL BERRENDO PENINSULAR (Antilocapra americana peninsularis)
POPULATION AND HABITAT VIABILITY ASSESSMENT FOR THE PENINSULAR PRONGHORN (Antilocapra americana peninsularis)
La Paz, Baja California Sur, Mexico Centro de Investigaciones Biologicas del Noroeste, S.C. 16-18 de Noviembre de 1994
Editado por Edited by
Jorge Cancino, Philip Miller, Jose Bernal Stoopen, and Jim Lewis
Recopilado por los Participantes del Taller Compiled by the Workshop Participants
Un Taller en Colaboraci6n: A Collaborative Workshop of
Centro de Investigaci6n Biol6gicas del Noroeste, S.C. (CIB) Secretaria de Desarrollo Social (SEDESOL) IUCN/SSC Conservation Breeding Specialist Group
Ci:l SPECIES SuRVIVAL COMMISSION ~ SEDESOL A contribution of the IUCN/SSC Conservation Breeding Specialist Group.
Cover Photo: El Berrendo Peninsular (Antilocapra americana peninsularis) Desierto de Vizcaino, Baja California Sur, Mexico Photo by Patricio Robles Gil, Agrupaci6n Sierra Madre, Mexico.
Cancino, J., P. Miller, J. Bernal Stoopen, and J. Lewis (eds.). 1995. Population and Habitat Viability Assessment for the Peninsular Pronghorn (Antilocapra americana peninsularis). IUCN/SSC Conservation Breeding Specialist Group: Apple Valley, MN. 113pp. The work of the Conservation Breeding Specialist Group is made possible by generous contributions from the following members of the CBSG Institutional Conservation Council
Conservators ($10,000 and above) North of England Zoological Society, Union of German Zoo Directors Chester Zoo Wellington Zoo Australasian Species Management Prog. Oklahoma City Zoo World Parrot Trust Chicago Zoological Society Paignton Zoological & Botanical Gardens Zoo de Ia Casa de Campo-Madrid Columbus Zoological Gardens Parco Natura Viva Garda Welsh Mt. Zoo/Zoo!. Society of Wales Denver Zoological Gardens Zoological Park Zoologischer Garten Frankfurt Fossil Rim Wildlife Center Penscynor Wildlife Park Friends of Zoo Atlanta Philadelphia Zoological Garden Curators ($250-$499) Greater Los Angeles Zoo Association Phoenix Zoo International Union of Directors of Pittsburgh Zoo Camperdown Wildlife Center Zoological Gardens Riverbanks Zoological Park Emporia Zoo Metropolitan Toronto Zoo Royal Zoological Society of Antwerp Edward D. Plotka Minnesota Zoological Garden Royal Zoological Society of Scotland Racine Zoological Society Omaha's Henry Doorly Zoo San Francisco Zoo Roger Williams Zoo Saint Louis Zoo Schoenbrunner Tiergarten Thrigby Hall Wildlife Gardens Sea World, Inc. Sedgwick County Zoo Topeka Zoological Park White Oak Conservation Center Sunset Zoo (10 year commitment) Tropical Bird Garden Wildlife Conservation Society - NY Taipei Zoo Zoological Society of Cincinnati The WILDS Sponsors ($50-$249) Zoological Society of San Diego The Zoo, Gulf Breeze, FL Urban Council of Hong Kong African Safari Guardians ($5,()()()..$9,999) Washington Park Zoo Apenheul Zoo Wassenaar Wildlife Breeding Centre Belize Zoo Cleveland Zoological Society Wilhelma Zoological Garden Claws 'n Paws John G. Shedd Aquarium Woodland Park Zoo Darmstadt Zoo Loro Parque Y ong-In Farmland Dreher Park Zoo Lubee Foundation Zoological Parks Board of Victoria Fota Wildlife Park North Carolina Zoological Park Zoological Park Organization Great Plains Zoo Toledo Zoological Society Zoological Society of London Hancock House Publisher Wild Animal Habitat Zurich Zoological Garden Kew Royal Botanic Gardens Zoological Parks Board of Lisbon Zoo New South Wales Stewards ($500-$999) Miller Park Zoo Nagoya Aquarium Protectors ($1,00()-$4,999) Aalborg Zoo National Audubon Society-Research Arizona-Sonora Desert Museum Ranch Sanctuary Africam Safari Banham Zoo National Aviary in Pittsburgh Audubon Institute Cotswold Wildlife Park Ocean World Taipei Incorporation Bristol Zoo Dutch Federation of Zoological Gardens PAAZAB Burgers' Zoo Erie Zoological Park Parco Faunistico "La Torbiera" Caldwell Zoo Fota Wildlife Park Potter Park Zoo Calgary Zoo Givskud Zoo Tenerife Zoo Cologne Zoo Granby Zoological Society Tokyo Zoological Park Copenhagen Zoo International Zoo Veterinary Group Touro Pare-France Detroit Zoological Park Knoxville Zoo El Paso Zoo Lincoln Park Zoo Supporters ($25-$49) Federation of Zoological Gardens of Great National Geographic Magazine Britain and Ireland National Zoological Gardens of South Alameda Park Zoo Fort Wayne Zoological Society Africa American Loriinae Conservancy Fort Worth Zoo Odense Zoo Bighorn Institute Gladys Porter Zoo Orana Park Wildlife Trust Brandywine Zoo Houston Zoological Garden Paradise Park DGHT Arbeitsgruppe Anuren Indianapolis Zoological Society Perth Zoological Gardens Folsom Children's Zoo & Botanical International Aviculturists Society Porter Charitable Trust Garden Japanese Association of Zoological Parks Rolling Hills Ranch (5 year commitment) International Crane Foundation & Aquariums Rostock Zoo Jardin aux Oiseaux Jersey Wildlife Preservation Trust Royal Zoological Society of Lee Richardson Zoo Living Desert Southern Australia Natal Parks Board Marwell Zoological Park Rotterdam Zoo Oglebay's Good Children's Zoo Milwaukee County Zoo Tierpark Rheine Speedwell Bird Sanctuary NOAHS Center Twycross Zoo 1 August 1995
CONTENIDO CONTENTS
Introduccion - Introduction Xl
Seccion 1: Resumen Ejecutivo 3 Executive Summary 5
Seccion 2: Biologia de Ia Poblacion y Modelos 9
Introducci6n 9 Panimetros de Entrada para la Simulaci6n 9 Resultados del Modelaje de Simulaci6n 12 Conclusiones 15 Muestra de un archivo de entrada para VORTEX 24 Tablas 25 Leyenda de Figuras 33 Figuras 35
Population Biology and Modelling 17 Introduction 17 Input Parameters for Simulations 17 Results from Simulation Modeling 20 Conclusions 22 Sample VORTEX Input File 24 Tables 25 Figure Legends 33 Figures 35
Seccion 3: Distribucion y Estado de Conservacion 43
Distribuci6n Hist6rica 43 Distribuci6n Actual 43 Lineas de Investigaci6n 43 Estado de Conservaci6n del Berrendo 44
Distribution and Conservation Status 47 Historical Distribution 47 Current Distribution 47 Research Needs 47 Conservation Status of the Pronghorn 48
Peninsular Pronghorn PHVA Report vii CONTENIDO (Continuacion) CONTENTS (Continued)
Seccion 4: Calidad de Habitat y Tendencias 51 Antecedentes 51 Necesidades de Investigaci6n 51
Habitat Quality and Trends 53 Background 53 Research Needs 53
Seccion 5: Amenazas 57 Transformaci6n y Reducci6n de Habitat 57 Caceria Ilegal 57 Ganaderia Extensiva 57 Factores Naturales 58 Conclusiones 58 Acciones 58
Threats 59 Transformation and reduction of habitat 59 Poaching 59 Extensive Livestock Production 59 Natural Factors 60 Conclusions 60 Actions 60
Seccion 6: Educacion Ambiental 63 Campafias de Difusi6n 63 Nivel Local, Regional, Nacional e Intemacional 64 Acciones a Corto Plazo 64
Environmental Education 67 Advertising Campaign 67 Local, Regional, National and International 68 Short-term Actions 68
Seccion 7: Investigacion- Research 73
Seccion 8: Estrategias de Manejo 77 Manejo en Cautiverio y Semicautiverio 77 Programa de Vigilancia 77 Normatividad del Uso del Suelo del Area de Distribuci6n del Berrendo 77 viii Peninsular Pronghorn PHVA Report CONTENIDO (Continuacion) CONTENTS (Continued)
Sefializaci6n 77 Tenencia de la Tierra y Uso del Suelo 77 Programa de Control de Depredadores 77 Manejo de Habitat 78
Management Strategies 79 Captive and Semi-Captive Management 79 Law Enforcement 79 Land Use Regulation of Present Habitat 79 Radio Telemetry 79 Type ofLand and Soil Use 79 Predator Control Programs 79 Habitat Management 80
Seccion 9: Referencias 83 References
Seccion 10: Lista de Participantes 87 List of Participants
Seccion 11: Referenda Tecnica de VORTEX 93 VORTEX Technical Reference
Peninsular Pronghorn PHVA Report ix
La p~, Baja California Sur:, Jueves 13 de Octubre de _1 994 Se -Reuniran AquFinvest~gadores Para . Analizar el Caso del Berrendo:JBA · ·Arturo Nieves Ramos llnos 30 invesllgadores de animales en pcligro de extinci6n. de ..Mexico y los Estados Unidos. sc' rcuniran-er1 es.taciudad a mediados de noviemhrc. para analizar Ia situaci6n del Berrendo Peninsular. du rante el Taller Apalisis de Ia Viabilidad del Habitat de Ia subcspecie sudcalifomiana. . . . El evento tcndra Iugar en las instalaciones del Centro' de lnvestiga- 8-A- ciones Biol6gicas del Noroeste. del 16 al 18 'del mes entrante, infor m6 Alfredo Berrn~dez Almada, Subdelegado de Ecologfa de Ia Sedesol. Durante los trabajos se analizani Ia situaci6n especffica del berren do peninsular,· principalmente lo relativo a las causas de disminuci6n de Ia poblaci6n, entre las que se incluyen Ia creaci6n de huevos cen tres urbanos, Ia caza furtiva y los depredadores naturales. Asimismo, se abordanin aspectos sabre medidas para evitar Ia ex tinci6n de Ia subespecie unica en el mundo, que de acuerdo a los cen sos efectuados en. los ultimos tiempos, fluctuan entre las 80 y 100 unidades. Este taller es organizado _porIa Delegaci6n Estatal de Ia Secretarfa de Desarrollo Social; conjuntamente con el CIB-NOR, el Instituto Nacional de Ecologfa, Ia empresa Exportadora de Sal y Ia Secretarfa de Agricultura y Recursos Hidniulicos. · Asistinin representantes de dependencias, basicamente de Ia SARH y SEDESOL, asf como las <;lependencias hom6nimas de los Estados Unidos, ademas de universidades e instituciones de investi- gacion de ambas naciones. •
Para ello se han girado unas 40 invitaciones a reconocidos espe cialistas en Ia materia, sin embargo se estima que seran unos 30 los que acudan, pues actualmente ya han confirrnado asistencia aproxi madamente ~0 personas.
Berrnuaez Almada apunt6 que para Ia entidad este evento sera de gran importancia, puesto que de el podrfan desprenderse actividades simi lares hacia otras especies sudcalifornianas, que se encuentran amenazadas de desaparecer. Entre estas se puede mencionar a El Aguila Real, el Aguila Pesca dorc, Coyotes, zorras y Ia tortuga caguama, esta ultima que a nivel nacional recibe especial atenci6n.
Peninsular Pronghorn PHVA Report xi
EVALUACION DE VIABILIDAD DE LA POBLACION Y DEL HABITAT DEL BERRENDO PENINSULAR (Antilocapra americana peninsularis)
POPULATION AND HABITAT VIABILITY ASSESSMENT FOR THE PENINSULAR PRONGHORN (Antilocapra americana peninsularis)
La Paz, Baja California Sur, Mexico Centro de Investigaciones Biologicas del Noroeste, S.C. 16-18 de Noviembre de 1994
SECCION 1
RESUMEN EJECUTIVO EXECUTIVE SUMMARY
RESUMEN EJECUTIVO
Historicamente la poblacion del berrendo peninsular (Antilocapra americana peninsularis) ocupaba aproximadamente 40,000 km2 desde San Felipe y San Quintin en el norte hasta Bahia Magdalena en e1 sur. En 1925 se estimo que el tamafio de la poblacion era de 500 individuos. Entre 1950 y 1980 debido principalmente al impacto de actividades humanas esta distribucion y abundancia n'ipidamente declino hasta niveles muy criticos: en la actualidad solamente de 100- 200 individuos se distribuyen en 3600 km2 aproximadamente. Este pobiacion se localiza unicamente en el Desierto de Vizcaino.
Este Taller de Am'ilisis de la Viabilidad de la Poblacion y del Habitat del berrendo peninsular se propuso para analizar la situacion con un grupo de especialistas representantes de diversas instituciones mexicanas y norteamericanas, tanto gubemamentales como no gubemamentales, siendo este Taller el primer evento de esta naturaleza que se realiza en Mexico con la participacion del CBSG/IUCN.
Participaron un total de 30 personas de 17 instituciones. El evento se realizo del 16 al 18 de noviembre de 1994 en el Centro de Investigaciones Biologicas del Noroeste, S. C. situado en La Paz, Baja California Sur, Mexico.
Se proporciono una extensa explicacion sobre el programa VORTEX y sus potencialidades.
Las metas que se marcaron para ser cubiertas por este taller fueron:
1. Elaboracion de un programa de manejo para la recuperacion del berrendo peninsular.
2. Identificacion de las instituciones (presentes y potenciales) con interes y compromiso en el programa de manejo para la recuperacion.
3. Definicion de las necesidades y prioridades para la recuperacion de esta subespecie.
4. Reunir y compilar la informacion basica y fundamentos teoricos, de los factores biologicos, no biologicos y sociales que influyen sobre la poblacion del berrendo peninsular.
Se formaron siete subgrupos de trabajo que abordaron los siguientes temas: biologia basica y modelaje; distribucion y estado; factores de riesgo; calidad del habitat y tendencias; y estrategias de conservacion ( educacion ambiental, investigacion, y manejo ). Se generaron los siete reportes respectivos, que ilustran:
Peninsular Pronghorn PHVA Report 3 a. El modelaje con VORTEX de la poblaci6n del berrendo peninsular indic6 que la poblaci6n es muy sensible al incremento de la mortalidad de crias, y que la mortalidad de adultos tambien es un factor importante. b. La distribuci6n de la subespecie se redujo hist6ricamente en mas de un 90%. El berrendo peninsular esta considerado, nacional e internacionalmente, en peligro de extinci6n. c. Las amenazas identificadas incluyen el furtivismo, la transformaci6n y reducci6n del habitat y la alta mortalidad de crias y juveniles. d. La calidad del habitat se considera de una baja capacidad de carga que se refleja en una baja densidad.
En cuanto a las estrategias de conservaci6n se propuso: e. I. Realizar una campafia de educaci6n ambiental a niveles local, estatal, nacional e internacional, que involucra a organizaciones gubernamentales y no gubemamentales. Algunas de las acciones que se sefialaron ya estan iniciadas. e.2. Se sefialaron las prioridades de investigaci6n en lo que se refiere ala poblaci6n destacando: biologia basica, monitoreo y uso del habitat. e.3. De manera similar al punto anterior se priorizaron las medidas a tomar, de las cuales las principales son: presencia en el area, gesti6n interinstitucional, analisis y reglamentaci6n del uso del suelo, monitoreo de la poblaci6n y crianza en semicautiverio.
Se definieron temporalidades para el inicio y/o continuaci6n de los diferentes rubros mencionados, asi como tambien se definieron las prioridades, responsabilidades e instituciones participantes y de apoyo.
4 Peninsular Pronghorn PHVA Report EXECUTIVE SUMMARY
Historically, the peninsular pronghorn (Antilocapra americana peninsularis) population occupied approximately 40,000 km2 from San Felipe and San Quintin in the north to Magdalena Bay in the south. In 1925 the population contained an estimated 500 individuals. Between 1950 and 1980, due principally to the impact of human activities, population distribution and abundance rapidly declined to very critical levels: presently, about 100-200 individuals are 2 distributed within approximately 3,600 km . This population is localized primarily within a portion of the Vizcaino Desert.
A Population Habitat and Viability Assessment (PHVA) Workshop was conducted for the peninsular pronghorn to analyze the current situation. A group of specialists representing various governmental and nongovernment institutions from Mexico and the United States participated in the Workshop. This was the first such workshop developed in Mexico with the participation of IUCN/CBSG.
Thirty persons from 17 institutions participated in the workshop held November 16 to 18, 1994, at the Center for Biological Research (N.W.), S.C. in La Paz, Baja California Sur, Mexico. Participants were given an extensive explanation of the VORTEX population simulation software package and its uses.
Goals identified by Workshop participants were:
1. To develop a management plan for recovery of the peninsular pronghorn.
2. To identify the institutions (present and potential), their interests and commitment to the recovery program.
3. To define the need and priorities for recovery of the subspecies.
4. To collect and record basic and fundamental information on social, biological, and nonbiological factors.
Working groups were organized on the following topics: life history/modelling; distribution and status; threats; habitat quality and trends; conservation education; research; and general management. The following were among the key points discussed in these groups: a. VORTEX modelling of the pronghorn population indicated that the population was very sensitive to increases in fawn mortality, with adult mortality also an important factor. b. Distribution of the subspecies has been reduced by more that 90%. The peninsular pronghorn has been declared an endangered species nationally and internationally.
Peninsular Pronghorn PHVA Report 5 c. Threats to the population include poaching, habitat loss, habitat quality reduction, and high mortality of both fawns and yearlings. d. The habitat quality is believed to be reduced in carrying capacity and it is reflected in the low pronghorn density.
Regarding conservation strategies, the participants: e 1. Proposed to develop an environmental education campaign at local, state, national, and international levels with the participation of government and nongovernmental organizations. Some of the actions that were identified are currently being developed. e2. Identified research priorities for the population considering mainly basic biology and monitoring of the subspecies and land use. e3. Prioritized recovery measures, the most important being: law enforcement in the area, interinstitutional analysis and regulation of the use of the land, monitoring of the population, and rearing in semi captivity.
A schedule was identified to initiate and continue recovery activities. Work priorities and responsibilities were defined and the role of participating institutions identified (e.g., funding or legal support).
6 Peninsular Pronghorn PHVA Report EVALUACION DE VIABILIDAD DE LA POBLACION Y DEL HABITAT DEL BERRENDO PENINSULAR (Antilocapra americana peninsularis)
POPULATION AND HABITAT VIABILITY ASSESSMENT FOR THE PENINSULAR PRONGHORN (Antilocapra americana peninsularis)
La Paz, Baja California Sur, Mexico Centro de Investigaciones Biologicas del Noroeste, S.C. 16-18 de Noviembre de 1994
SECCION2
BIOLOGIA DE LA POBLACION Y MODELOS POPULATION BIOLOGY AND MODELLING
BIOLOGIA DE LA POBLACION Y MODELAJE
Introducci6n
La necesidad de las estrategias intensivas del manejo y los efectos de estas pueden ser modeladas para suguerir que pnicticas pueden ser las mas efectivas para conservar esta poblacion. VORTEX, un programa de simulacion de modelaje escrito por Robert Lacy y Kim Hughes, fue utilizado como una herramienta para estudiar la interaccion de varias variables estocasticas.
El programa VORTEXes una simulacion Monte Carlo sobre los efectos tanto de las fuerzas deterministicas, como de los eventos demogra±icos, ambientales y geneticos estocasticos en las poblaciones silvestres. VORTEX modela las dinamicas de las poblaciones como eventos discretos y secuenciales (por ejemplo, nacimientos, muertes, relacion de sexos, catastrofes etc.) que ocurren de acuerdo a probabilidades definidas de distribucion. Las probabilidades de eventos son modeladas como variables constantes o al azar que siguen distribuciones especificas. El paquete simula una poblacion considerando una serie de eventos que describen el ciclo de vida tipico de organismos diploides con reproduccion sexual.
VORTEX no pretende dar respuestas absolutas. debido a que proyecta estocasticamente las interacciones de los varios parametros que entran en el modelo y debido tambien a los procesos aleatorios que ocurren en la naturaleza. La interpretacion de los resultados depende de nuestro conocimiento de la biologia del berrendo, de las condiciones que afectan ala poblacion y, de los posibles cambios que pueden ocurrir en el futuro. Mas especificamente, debido a que la cantidad de informacion sobre la historia de vida especifica para el berrendo peninsular es limitada, utilizamos informacion de la poblacion silvestre del berrendo de Sonora (Antilocapra americana sonorensis) cuando fue apropiado. Estas dos subespecies son ecologicamente similares, hecho que permite utilizar informacion de esta poblacion mayormente estudiada.
Parametros de Entrada para Ia Simulaci6n
Sistema de Apareamiento : Poligamos.
Edad Promedio a Primera Reproduccion: VORTEX define ala reproduccion como la edad en que las crias nacen, no la edad de madurez sexual. La edad en que las hembras entran en madurez sexual es de alrededor de 16 meses, mientras que los machos alcanzan la madurez sexual antes de cumplir un afio de vida. Sin embargo, debido a que el periodo de gestacion para esta subespecie es de aproximadamente 8 meses, determinamos la edad de primera reproduccion para las hembras a los dos afios y para los machos al primer afio de vida.
Edad de Senectud: VORTEX asume que los animales pueden reproducirse (en una tasa normal) a lo largo de su vida adulta. Informacion del berrendo de Sonora indica que tanto los machos como las hembras cuentan con capacidad reproductiva hasta los 10 anos.
Peninsular Pronghorn PHVA Report 9 Produccion de Crias: Con base a observaciones de campo para el berrendo de Sonora, estimamos que aproximadamente el 25% de las hembras con capacidad reproductiva no producen crias cada ano. Asi mismo, debido a que las hembras tipicamente tienen una sola cria durante su primera gestacion, subdividimos el tamafio de la camada de tal forma que el 7.5% de la poblacion de hembras tuviera una cria y el67.5% tuviera dos crias. La variacion en la reproduccion se modela en VORTEX introduciendo una desviacion standard (DS) para la proporcion de hembras adultas que no producen crias. Debido a la falta de datos empiricos con respecto a esta variable, asumimos que esta variacion (resultante de las fluctuaciones en la abundancia de alimentos y en la variacion en las que las hembras alcanzan la madurez sexual) fue de 25% del valor de la media. VORTEX entonces determina el porcentaje de hembras que se reproducen cada afio de la simulacion, a partir de un distribucion binomial con la media especificada (25%) y laDS (6.25%). Las proporciones relativas de las camadas de 1 y 2 crias se mantienen constantes.
Machos en el Pool Reproductive: La proporcion de machos con capacidad reproductiva anual inicialmente fue establecida en un 50%. Debido a que existe unajerarquia para los machos reproductivos, creemos que los machos mas maduros conforman el pool reproductive. Debido a que las densidades de las poblaciones en los desiertos son bajas, utilizamos porcentajes mayores de machos reproductivos que los que han sido reportados en la literatura debido al intervalo de tiempo que requiren los machos para encontrar hembras. Nosotros establecimos un porcentaje mayor correspondiente al 75%.
Relacion de Sexos al Nacimiento: Debido a que no existe informacion para el berrendo peninsular que indique una relacion de sexos en el nacimiento diferente a 50:50 utilizamos una relacion igual para todas las simulaciones.
Mortalidad: La sobrevivencia de las crias entre las poblaciones de berrendos de Norteamerica pueden ser influenciada altamente por depredacion, principalmente por depredadores tales como los coyotes y gato montes. Por ejemplo, la mortalidad en crias debido a la depredacion por coyote en Alberta, Canada fue de cerca del 50% (Barrett, 1984). Este nivel de mortalidad ocurrio en una area con relativamente baja densidad de coyotes, estimada 2 conservativamente en un ejemplar por cada 13 km • Las investigaciones desarrolladas en el Desierto del Vizcaino, sugieren que las densidades de coyotes pueden ser mucho mayores. En forma consecuente, inicialmente fijamos una tasa de mortalidad en juveniles correspondientes al 60%. Esta decision estuvo tambien basada en los resultados de las relaciones de sexo y edad reportados en los mas recientes rastreos del berrendo peninsular. Asi mismo, los datos de las poblaciones de berrendos Americanos en los zoologicos indican que la mortalidad en juveniles es de cerca del 50%, por lo que la estimacion de un 60% de mortalidad en vida silvestre parece muy razonable. Por supuesto, esta discusion no implica que toda la mortalidad en el berrendo se debe a depredacion. Muchos otros factores juegan un papel en la mortalidad natural. Nosotros investigamos un aumento adicional en la densidad de depredadores incrementando la mortalidad juvenil a un 70% y hasta a un 80%.
10 Peninsular Pronghorn PHVA Report No existen datos especfficos de mortalidad en adultos para el berrendo Peninsular, pero datos provenientes de las poblaciones del berrendo Sonorense indican que esta es aproximadamente del 10% para ambos sexos. Para investigar los efectos del aumento en la depredacion y en la caceria ilegal en el componente adulto de la poblacion, establecimos valores de mortalidad adulta en un 15% y 20% en una submuestra de la simulaciones.
Depresion por Consanguinidad: No existe informacion especifica de la prevalencia y de los efectos de la consangunidad en las poblaciones de berrendos peninsulares. De acuerdo a datos de censos recientes, puede ser razonable el inferir que el reducido tamano de las poblaciones observadas recientemente pueda ser resultado de un grado medible de consanguinidad. Por lo tanto, hemos incluido a la depresion por consanguinidad en una submuestra de los escenarios de modelaje.
Utilizamos el modelo de heterosis de depresion por consanguinidad. En este modelo los individuos que son heterozigotos para un locus genetico particular tienen un vigor superior que aquellos que son homocigotos para este locus. Debido a que los alelos detrimentales particulares no son removidos en el curso del tiempo por la seleccion natural en este modelo, puede ser que los efectos deletereos de la consanguinidad en las poblaciones de berrendos anteriormente modeladas esten sobreetimados.
La severidad de depresion por consanguinidad en las poblaciones de mamiferos puede ser medida como el nlimero de "equivalentes letales" contenidos en el genoma de la poblacion de interes. Informacion de algunas especies de cervidos en cautiverio, sugiere que estas especies albergan alrededor de 3 equivalentes letales, un valor muy cercano al de la mediana obtenido en una muestra mayor que incluye a 40 especies de mamiferos analizada por Ralls et al. (1988). En forma consecuente, modelamos la depresion por consanguinidad utilizando el valor de la mediana de equivalentes letales.
Tamafio de la Poblacion Inicial: El tamafio de la poblacion fue inicialmente establecido en 100 animales. Este valor fue primordialmente seleccionado con base al juicio profesional, debido a la considerable variacion de los censos obtenidos en los afios recientes. Es importante el senalar que los ultimos dos censos indicaron una cuenta minima de entre 51 y 17 5 animales, correspondiendo el valor menor al censo mas reciente (Abril, 1994). El valor de 100 noes una media para estos dos censos. El promedio de los dos ultimos censos es mayor pero los grupos alcanzaron un concenso de que la poblacion real estaba constituida por alrededor de 100 individuos.
Sin embargo, es posible que las tecnicas de muestreo utilizadas en el censo, revelaron unicamente una proporcion de la poblacion real. Por lo tanto, empezamos un numero de simulaciones considerando 200 animales. Iniciamos todos los escenarios con una distribucion estable de edades, que distribuye a la poblacion inicial entre cada una de las clases por edad y sexo, de acuerdo con la mortalidad existente y la calendarizacion reproductiva.
Peninsular Pronghorn PHVA Report 11 Capacidad de Carga: La capacidad de carga (K) define un limite superior para el tamaiio de la poblacion, arriba del cual se impone una mortalidad equitativa a traves de las clases de edad y sexo, regresando la poblacion a este valor o limite.
Nelson (1925) indico que el tamaiio total de la poblacion del berrendo peninsular a lo largo de la peninsula, con base en metodos de censo disponible entonces, era de 500 individuos. La distribucion actual del berrendo, dentro de la Reserva del Desierto del Vizacaino, incluye unicamente una parte de la distribucion historica. Debido a que los nfuneros reportados por Nelson probablemente subestimaron del tamafio real de la poblacion y, debido ala superficie del habitat de la Reserva de la Biosfera, nosotros estimamos una capacidad de carga de 500 individuos para la distribucion actual. Debido a que casino existe informacion sobre la capacidad de carga en la region para el berrendo, esta estimacion esta basada principalmnte en la intuicion disponible. Debido a que la distribuci6n actual del berrendo se encuentra dentro de la Reserva de la Biosfera, no modelamos una reducci6n deterministica en la disponibilidad del habitat ni tampoco la capacidad de carga del berrendo resultante de la degradaci6n del habitat por motivos humanos.
Catastrofes: Las catastrofes son eventos singulares que ocurren fuera de los limites de la variaci6n ambiental normal que afecta la reproducci6n y sobrevivencia. Estos eventos pueden ser huracanes, inundaciones, sequias, incendios, enfermedades y otras circunstancas similares. Los catastrofes son modeladas asignando una probabilidad anual de ocurrencia y un factor de severidad con un rango de 0.0 (efecto maximo o absoluto) a 1.0 (sin efecto).
Nosotros incluimos un 100 - aiio de sequia en un conjunto de escenarios. Este evento por lo tanto tiene una probabilidad de ocurrencia anual del 1.0% y reduce la probabilidad de sobrevivencia de todos los individuos en un 50% en comparaci6n con aquellos afios en los que no se registra ninguna sequia.
Repeticiones y Afios de Proyecci6n: Los escenarios que no incluyen depresi6n por consanguinidad fueron simulados 500 veces y aquellos que incorporaron depresi6n por consanguinidad fueron simulados 100 veces, debido a limitantes de computaci6n. Las proyecciones de las poblaciones fueron extendidas a 100 afios. Los resultados fueron resumidos en peri6dos o intervalos de 10 afios para su utilizaci6n en figuras y tab las que mostramos a continuaci6n. Todas las simulaciones fueron desarrolladas utilizando el paquete software VORTEX6.3.
Resultados del Modelaje de Simulacion
Resultados Deterministicos
Tasa de crecimiento (rct} Las tasas de crecimiento deterministicas calculadas utilizando los metodos de la matriz Leslie (Ricklefs 1979) se muestran para cada escenario en la columna 6 de
12 Peninsular Pronghorn PHVA Report las Tablas 1 - 8. Los valores positivos indican crecimiento de poblaciones, mientras que los valores negativos indican una disminucion de la poblacion. Un valor deterministico r que resulta de simulaciones (r5, referirse abajo) puede mostrar una indicacion del impacto de los factores estocasticos en la persistencia de las poblaciones. Estas tasas deterministicas de crecimiento se calculan a partir de la mortalidad y de la calendarizacion reproductiva para cada escenario de modelaje. Como resultado, parametros tales como la proporcion de machos adultos disponibles para reproduccion, el tarnafio inicial de la poblacion, y la inclusion de depresion por consanguinidad no alteran las tasas de crecimiento para una mortalidad particular. Este hecho es evidente en estas tablas. Sin embargo, observese que la inclusion de una catastrofe - en este caso, una sequia de 100 afios, si reduce la tasa deterministica de crecimiento, aunque solo ligeramente (Tablas 5 - 8). Independientemente de la severidad de la mortalidad adulta, las poblaciones siempre se encuentran en crecimiento deterministico cuando la mortalidad de las crias es del 60%. Sin embargo, la tasa de crecimiento se reduce casi al 75% cuando la mortalidad de los adultos es mayor, en comparacion con una reducida mortalidad adulta. Cuando la mortalidad de las crias es del 67.5%, las poblaciones se encuentran en una declinacion deterministica unicamente en aquellas condiciones en las que la mortalidad adulta es elevada. Tal y como se espera, las poblaciones se encuentran en declinacion deterministica cuando la mortalidad de las crias es alta (75%). En forma cualitativa, estas observaciones son similares cuando se adiciona una sequia al modelo. Baia Mortalidad en Crias (60%) Resultados del escenario de modelaje base- baja mortalidad de crfas, tamafio inicial de la poblacion de 100 animales, sin catastrofes y sin depresion por consanguinidad- se muestran en la Tabla 1 y en la Figura 1. Bajo condiciones de baja mortalidad adulta, el riesgo de extincion durante el periodo de tiempo de 100 afios es de virtualmente 0 (Fig. la), conforme la poblacion se incrementa rapidarnente hacia la capacidad de carga bajo un crecimiento del 5 - 9 % (Fig. 1b). Cuando la mortalidad adulta es alta, el riesgo de extincion de la poblacion es de unicamente 3% y la poblacion se incrementa en una tasa anual de entre 1.5% a casi el 50% de la capacidad de carga asumida (Fig. 1b). En este y en todos los escenarios subsecuentes, el incrementar la porporcion de machos adultos disponibles para reproduccion del valor base de 50 al 75% produce valores casi identicos. Esto es de esperarse, debido al sistema poligarno de esta subespecie. Por lo tanto, podemos concluir que la persistencia de la poblacion del berrendo es de cierta forma insensible a este parametro; consecuentemente, consideraremos unicamente aquellos escenarios utilizando el valor base de 50% en todas las discusiones que siguen. Peninsular Pronghorn PHVA Report 13 Si asumirnos un tamafio inicial de la poblaci6n de 200 individuos y una baja mortalidad de crias, el riesgo de extinci6n es de virtualmente cero bajo todos los escenarios de mortalidad (Tabla 3, Fig. 2a). Es mas, las poblaciones se incrementan de una forma muy similar a aquellas de los escenarios base. Si se incorpora la depresi6n por consanguinidad al modelo, el riesgo de extinci6n se incrementa considerablemente, a mas del 40%, si el tamafio inicial de la poblacion es reducido y la mortalidad adulta es alta (Tabla 2, Fig. la). Aun cuando el tamano inicial de la poblaci6n es de 200, el riesgo de extinci6n se incrementa a 13% si la poblaci6n esta sufriendo de depresi6n por consanguinidad (Tabla 4, Fig. 2a). Sin embargo, bajo condiciones de alta mortalidad adulta, independientemente del tamafio inicial de la poblaci6n, las poblaciones muestran un incremento inicial seguido por un decremento en su tamano original debido a la alta la mortalidad que ocurre por consanguinidad (Figs. 1b, 2b ). Es claro que, bajo estas condiciones, la depresi6n por consanguinidad tiene un efecto pronunciado en las dinamicas de esta poblaci6n. Sin embargo, es importante recordar, que el rnodelo de heterosis de depresi6n por consanguinidad utilizado en estas simulaciones puede sobreestimar los efectos deletereos de la consanguinidad (referirse a Panimetros de Entrada para Ia Simulacion). La adici6n de sequias a los modelos incrementa el riesgo de extinci6n y disminuye ligeramente los tamafios finales de la poblaci6n, particularmente bajo condiciones de una alta mortalidad de adultos y de depresi6n por consanguinidad (Tablas 5 y 6, Figs. 7 y 8). Aful cuando estos eventos tienen una baja probabilidad de ocurrencia anual, sus efectos pueden ser muy severos y pueden conducir a la inestabilidad en las poblaciones. Mortalidad Intermedia de Crias (67.5%) En todos los casos una alta mortalidad de crias conduce a una mayor inestabilidad de la poblaci6n. Cuando la mortalidad adulta es baja, el riesgo de extinci6n de la poblaci6n disminuye para ambos tamafios iniciales de la poblaci6n, aun en la presencia de depresi6n por consanguinidad (Tablas 1-4, Figs. 3a y 4a). El riesgo se incrementa considerablemente conforme la mortalidad adulta se incrementa. El riesgo se incrementa a casi 1.0 bajo condiciones de una alta mortalidad de adultos y de depresi6n por cosanguinidad, con una media en el tiempo de extinci6n de aproximadamente 50 a 60 afios. Los efectos de la depresi6n por consanguinidad son particulamente pronunciados cuando el tamafio inicial de la poblaci6n es bajo y la mortalidad de adultos es intermedia: el riesgo de extinci6n se incrementa de 9% a 63%, en escenarios excentos de sequias, y de entre 18% al 69% en aquellos escenarios que incorporan sequias. Cuando las poblaciones iniciales son grandes, los efectos de la consanguinidad en el riesgo de extinci6n no son tan pronunciados, aunque deben ser considerados. Si la depresi6n por consanguinidad y la sequia estan ausentes, las poblaciones continuan incrementandose en un 0.2 - 4.1% por afio para los niveles bajos e intermedios de mortalidad adulta. Sin embargo, una alta mortalidad adulta conduce a una declinaci6n consistente de la poblaci6n cercana al cero (Tablas 1 y 3, Figs. 3 y 4). El adicionar depresi6n por consanguinidad reduce la tasa de crecimiento estocastica para todos los niveles de mortalidad adulta; de hecho, bajo un nivel intermedio de mortalidad adulta, el crecimiento estocastico cambia de positivo a 14 Peninsular Pronghorn PHVA Report negativo (Figs. 3b y 4b ). El incluir el componente de sequias en el modelo tiende a aumentar estos efectos. Mortalidad Alta de Crias (75%) Cuando se incrementa la mortalidad de crias al 75%, ocurre una rapida destabilizaci6n de la poblaci6n. Aun bajo condiciones relativamente 6ptimas, como por ejemplo, de un tamafio inicial de la poblaci6n de 200 individuos, sin depresi6n por consanguinidad y sin sequfa, el riesgo de extinci6n es del24% con una media en el tiempo de extinci6n de 72 afios (Tabla 3, Fig. 6a). El riesgo se incrementa rapidamente a 100% al adicionarse sequias, depresi6n por consanguinidad y una mayor mortalidad de adultos (por ejemplo Fig. 12a). Bajo estas condiciones mayormente severas, la extinci6n ocurre en un peri6do de 30-40 afios. Los efectos de depresi6n por consanguinidad se pronuncian mayormente cuando la mortalidad en adultos es baja; estos efectos son mas pronunciados bajo condiciones de mayor mortalidad de adutos (Figs. Sa, 6a, 11a y 12a). Todos los escenarios muestran decrementos rapidos y consistentes en el tamafio de las poblaci6n, con tasas de crecimiento estocasticas que presentan un rango del-2% al-13%. AI igual que con las mediciones del riesgo de extinci6n, la depresi6n por consanguinidad muestra los mayores efectos cuando la mortalidad adulta es reducida. Conclusiones Los resultados del modelaje discutidos aqui representan una gran variedad de situaciones. El escenario mas optimistico, con una mortalidad adulta y de crias reducida, sin sequias o depresi6n por consanguinidad y con un tamafio inicial de la poblaci6n largo, presenta una tasa anual de crecimiento de casi el9% sin riesgo de extinci6n en los proximos 100 afios (Tabla 1, Fig. 1). En el otro extremo, el escenario mas pesimista produce una poblaci6n que se extinguira dentro de 70 afios, con una tasa anual de crecimiento de -13%. Es probable que actualmente la poblaci6n se encuentre intermedia entre estos dos extremos, con un riesgo considerable de extinci6n de entre los pr6ximos 100 afios. Los modelos indican que tanto la mortalidad de crias como la adulta parecen ser los factores primarios que influencian la variabilidad del berrendo peninsular en el Desierto de Vizacaino. De hecho, los resultados sugieren que la poblaci6n es igualmente sensible a incrementos en la mortalidad de crias como de adultos. Sin embargo, los esfuerzos del modelaje muestran claramente que una reducida mortalidad en crias, aun con una alta mortalidad de adultos, puede conducir a un crecimiento o estabilidad poblacional, mientras que una mortalidad alta en crias siempre resulta en una declinaci6n de la poblacion y en la inestabilidad de la misma (refererise a Tabla 1). Estos datos sugieren que las acciones de manejo deben de dirigirse especificamente hacia la reducci6n en la mortalidad del berrendo, posiblemente haciendo una enfasis en las crias. Parece ser que esto puede lograrse traves del control de depredadores y de la crianza bajo Peninsular Pronghorn PHVA Report 15 condiciones de semicautiverio. Para determinar el curso de acci6n mas apropiado, deben de realizarse amilisis de costo-beneficio muy detallados. Aim mas, la consecuencias deletereas de la consanguinidad pueden ser considerables, particularmente cuando las poblaciones de los berrendos son pequefias (por ejemplo de 100) y estos efectos deben de ser reconocidos y apreciados por aquellos responsables del manejo de la subespecie. Esfuerzos directos deben de desarrollarse para establecer y mantener una poblaci6n mayor en vida silvestre en orden de reducir el impacto de la depresi6n por consanguinidad. 16 Peninsular Pronghorn PHVA Report POPULATION BIOLOGY AND MODELLING Introduction The need for and effects of intensive management strategies can be modelled to suggest which practices may be the most effective in conserving this population. VORTEX, a simulation modelling package written by Robert Lacy and Kim Hughes, was used as a tool to study the interaction of multiple stochastic variables. The VORTEX package is a Monte Carlo simulation of the effects of deterministic forces as well as demographic, environmental, and genetic stochastic events on wild populations. VORTEX models population dynamics as discrete, sequential events (e.g., births, deaths, sex ratios, catastrophes, etc.) that occur according to defined probability distributions. The probabilities of events are modelled as constants or as random variables that follow specified distributions. The package simulates a population by stepping through the series of events that describe the typical life cycle of sexually reproducing, diploid organisms. VORTEX is not intended to give absolute answers, since it is projecting stochastically the interactions of the many parameters which enter into the model and because of the random processes involved in nature. Interpretation of the output depends upon our knowledge of the biology of the pronghorn, the conditions affecting the population, and possible changes in the future. More specifically, since the amount of life history data specific to the peninsular pronghorn is limited, we utilized data from wild population of Sonoran pronghorn (Antilocapra americana sonoriensis) when appropriate. These two subspecies are similar ecologically, allowing for the use of data from this well-studied population. Input Parameters for Simulations Mating System: Polygynous Average Age of First Reproduction: VORTEX defines breeding as the time when young are born, not the age of sexual maturity. Age of sexual maturity for females is about 16 months, while males reach sexual maturity before one year of age. However, because the gestation period for this subspecies is approximately 8 months, we have set the age of first reproduction for females at two years and for males at one year. Age of Senescence: VORTEX assumes that animals can breed (at the normal rate) throughout their adult life. Data from Sonoran pronghorn indicate that both males and females are reproductively capable until 10 years of age. Fawn Production: Based on field observations for Sonoran pronghorn we estimated that approximately 25% of reproductively capable females did not have fawns each year. Further, Peninsular Pronghorn PHVA Report 17 because does typically have single fawns during their first pregnancy we subdivided litter size so that 7.5% of the female population had one offspring and 67.5% had twin fawns. Variation in reproduction is modelled in VORTEX by entering a standard deviation (SD) for the proportion of adult females producing no offspring. Because empirical data for this variable were lacking, we assumed that such variation (due to fluctuations in food abundance and variability in the age at which females reach sexual maturity) was 25% of the mean value. VORTEX then determines the percent of females breeding each year of the simulation by sampling from a binomial distribution with the specified mean (25%) and SD (6.25%). The relative proportions of litters of 1 and 2 offspring are held constant. Male Breeding Pool: The proportion of adult males breeding each year was initially set at 50%. Because pronghorn have a social heirarchy for breeding males, we felt that only more mature bucks made up the breeding pool. Because densities in desert populations are low, we also used higher percentages of males actually breeding than reported in the literature due to the time interval required for bucks to find does. We set this higher figure at 75%. Sex Ratio at Birth: As no data exist indicating other than a 50:50 sex ratio at birth for peninsular pronghorns, we used an equal sex ratio for all simulations. Mortality: Fawn survival among pronghorn populations in North America can be heavily influenced by predation, primarily by such predators as coyotes and bobcats. For example, fawn mortality due to coyote predation in Alberta, Canada was nearly 50% (Barrett, 1984). This level of mortality was in an area of relatively low coyote density, estimated conservatively at one per 2 13 km • Research in the Vizcaino Desert suggests that coyote densities may be much higher. Consequently, we set an initial mortality rate for juveniles at 60%. This choice was also based on sex and age ratios from the most recent surveys of peninsular pronghorn. In addition, data from American pronghorn populations in zoos indicate juvenile mortalities approaching 50%, making 60% juvenile mortality in wild populations very reasonable. Of course, this discussion does not imply that all mortality in pronghorn is due to predation. Many other factors play a role in natural mortality. We modelled a further increase in predator density by increasing juvenile mortality to 70% and even 80%. No specific data are available for adult mortality of peninsular pronghorn, but data from Sonoran pronghorn populations indicate that adult mortality is about 10% for both sexes. To investigate the effects of increased predation and poaching on the adult component of the population, we set adult mortality at 15% and 20% in a subset of simulations. Inbreeding Depression: Specific data do not exist on the prevalence and effects of inbreeding in peninsular pronghorn populations. Given recent census data, it may be reasonable to infer that the small population sizes observed in the recent past may have resulted in some measurable degree of inbreeding. Therefore, we have included inbreeding depression in a subset of modelling scenarios. 18 Peninsular Pronghorn PHVA Report We employed the heterosis model of inbreeding depression, in which individuals that are heterozygous at a given genetic locus have superior fitness to those that are homozygous at that locus. Because particular detrimental alleles are not removed by natural selection from the population over time in this model, the heterosis model may provide an overestimate of the deleterious effects of inbreeding in the pronghorn populations modelled below. The severity of inbreeding depression in mammal populations can be measured as the number of "lethal equivalents" contained in the genome of the population of interest. Data for some captive cervid species suggests that these species harbor about 3 lethal equivalents, a value very close to the median value obtained in the larger dataset of 40 mammalian species analyzed by Ralls et al. (1988). Consequently, we modelled inbreeding depression using this median lethal equivalent value. Initial Population Size: The initial population size was first set at 100 animals. This value was selected based primarily on professional judgement due to considerable variation in census results in recent years. It is important to note that the last two censuses indicated between 51 and 175 animals minimum count with the lower value being the most recent (April1994). The value of 100 is not a mean for these two surveys. The average of the last two censuses is higher but the groups reached a concensus that the true population was closer to 100. It is possible, however, that the survey techniques used in the censuses revealed only a proportion of the true population. We therefore began a number of simulations with 200 animals. We initialized all scenarios with a stable age distribution that distributes the starting population among each age-sex class in accordance with the existing mortality and reproductive schedules. Carrying Capacity: The carrying capacity (K) defines an upper limit for the population size, above which additional mortality is imposed equally across age and sex classes in order to return the population to this value. Nelson (1925) stated that the total population size for peninsular pronghorn across the peninsula, based on census methods available at that time, was 500 individuals. The current distribution of the pronghorn, within the Vizcaino Desert Biosphere Reserve, encompasses only a part of the historical distribution. Since Nelson's number probably underestimated the true population size, and given the current size of Biosphere Reserve habitat, we estimated a carrying capacity of 500 individuals for the current distribution. As there are virtually no data that provide substantial insight into the pronghorn carrying capacity of the region, this estimate is based primarily on available intuition. Because the current distribution of pronghorn is within the Biosphere Reserve, we did not model a deterministic reduction in habitat availability and, by extension, pronghorn carrying capacity due to human-mediated habitat degradation. Peninsular Pronghorn PHVA Report 19 Catastrophes: Catastrophes are singular events outside the bounds of normal environmental variation affecting reproduction and survival. These events can be tornadoes, floods, droughts, fire, disesase, or other similar circumstances. Catastrophes are modelled by assigning an annual probability of occurrence and a severity factor ranging from 0.0 (maximum or absolute effect) to 1.0 (no effect). We included a 100-year drought in a set of scenarios. This event therefore has an annual probability of occurrence of 1.0% and reduces the probability of survival for all individuals by 50% compared to non-drought years. Iterations and Years of Projection: Those scenarios not including inbreeding depression were simulated 500 times, while scenarios incorporating inbreeding depression were simulated 100 times because of computational limitations. Population projections extended for 100 years. Output results were summarized at 10-year intervals for use in the Figures and Tables that follow. All simulations were conducted using the VORTEX 6.3 software package. Results from Simulation Modelling Deterministic Results Growth rate (rdl The deterministic growth rates calculated using Leslie matrix methods (Ricklefs 1979) are shown for each scenario in column 6 ofTables 1-8. Positive values indicate population growth, while negative values indicate population decline. A deterministic r < 0 is in deterministic decline (i.e., deaths outpace births), and will go extinct even in the absence of any stochastic fluctuations. The difference between the deterministic population growth rate and the , growth rate resulting from the simulation (r5 see below) can give an indication of the impact of stochastic factors on population persistence. These deterministic growth rates are calculated from the mortality and fecundity schedules for each modelling scenario. As a result, parameters such as the proportion of adult males available for breeding, the initial population size, and the inclusion of inbreeding depression do not alter the growth rates for a particular mortality schedule. This is evident in the Tables. Note, however, that the inclusion of a catastrophe-in this case, a 100-year drought-does reduce the deterministic growth rate, although only slightly (Tables 5-8). Regardless of the severity of adult mortality, the populations are always in deterministic growth when fawn mortality is 60%. However, the growth rate is reduced by nearly 75% when adult mortality is highest, compared to conditions of low adult mortality. When fawn mortaltiy is 67.5%, populations are in deterministic decline only under conditions of high adult mortality. As may be expected, the populations are in deterministic decline when fawn mortality is high (75%). These observations are qualitatively the same when drought is added to the models. 20 Peninsular Pronghorn PHVA Report Stochastic Simulation Results Low Fawn Mortality (60%) Results from the base modelling scenario-low fawn mortality, initial population size of 100 animals, with no catastrophes and no inbreeding depression-are shown in Table 1 and Figure 1. Under conditions oflow adult mortality, risk of extinction over the 100-year time span is virtually zero (Fig. 1a), as the population rapidly increases to carrying capacity under 5-9% annual growth (Fig. 1b). When adult mortality is high, the risk of population extinction is only 3% and the population increases at an annual rate of about 1.5% to nearly 50% ofthe assumed carrying capacity (Fig. 1b). In these and all subsequent scenarios, increasing the proportion of adult males available for breeding from the base value of 50% to 75% yields nearly identical results. This is to be expected, given the polygynous breeding system in this subspecies. We can therefore conclude that the persistence of the pronghorn population is quite insensitive to this parameter; consequently, we will consider only those scenarios using the base 50% value in all discussions to follow. If we assume an initial population size of200 individuals and low fawn mortality, the extinction risk is virtually zero under all adult mortality scenarios (Table 3, Fig. 2a). Furthermore, the populations increase in a manner very similar to those in the base scenarios. If inbreeding depression is incorporated into the model, extinction risk is increased considerably, to over 40%, if the initial population size is low and adult mortality is high (Table 2, Fig. 1a). Even if the initial population size is 200, the risk of extinction increases to 13% if the population is suffering from inbreeding depression (Table 4, Fig. 2a). However, under conditions of high adult mortality, regardless of initial size, the populations show an initial increase followed by a decrease to their original size as increased mortality occurs through inbreeding (Figs. 1b, 2b ). It is clear that, under these conditions, inbreeding depression does have a pronounced effect on the dynamics of this population. It is important to remember, however, that the heterosis model of inbreeding depression used in these simulations may provide an overestimate of the deleterious effects of inbreeding (see Input Parameters for Simulations above). The addition of drought to the models increases the extinction risk and slightly decreases the final population sizes, particularly under conditions of high adult mortality and inbreeding depression (Tables 5 and 6, Figs. 7 and 8). Even though these events have a low annual probability of occurrence, their effects can be quite severe and can lead to population instability. Intermediate Fawn Mortality (67.5%) Overall, higher fawn mortality leads to greater population instability. When adult mortality is low, the risk of population extinction is negligible for both initial population sizes, even in the presence of inbreeding depression (Tables 1-4, Figs. 3a and 4a). This risk is greatly increased, however, as adult mortality is increased. This risk increases to nearly 1.0 under conditions of Peninsular Pronghorn PHVA Report 21 high adult mortality and inbreeding depression, with a mean time to extinction of approximately 50-60 years. The effects of inbreeding depression are particularly pronounced when the initial population size is low and adult mortality is intermediate: the risk of extinction increases from 9% to 63% in non-drought scenarios, and from 18% to 69% in scenarios incorporating drought. When initial populations are larger in size, the effects of inbreeding on extinction risk are not as pronounced, but are nevertheless considerable. If inbreeding depression and drought are absent, populations continue to increase at 0.2-4.1% per year for low and intermediate levels of adult mortality. However, high adult mortality leads to consistent population decline to near zero (Tables 1 and 3, Figs. 3 and 4). Adding inbreeding depression further reduces the stochastic growth rate for all levels of adult mortality; in fact, under intermediate adult mortality, stochastic growth changes from positive to negative (Figs. 3b and 4b). Including drought into the models tends to magnify these effects. High Fawn Mortality (75%) When fawn mortality is further increased to 75%, rapid population destabilization occurs. Even under relatively optimal conditions, i.e., initial population size of200 individuals, no inbreeding depression, and no drought, the risk of extinction is 24% with a mean time to extinction of 72 years (Table 3, Fig. 6a). This risk increases rapidly to 100% as drought, inbreeding depression, and higher adult mortality are factored into the models (i.e., Fig. 12a). Under these more severe conditions, extinction occurs in 30-40 years. The effects of inbreeding depression are most pronounced when adult mortality is low; these effects are swamped out under conditions of higher adult mortality (Figs. 5a, 6a, 11a, and 12a). All scenarios show consistent, and in many cases rapid, decline in population size, with stochastic growth rates ranging from -2% to -13%. As with measures of extinction risk, inbreeding depression shows the greatest effect when adult mortality is low. Conclusions The modelling results discussed here represent a great variety of outcomes. The most optimistic scenario, with low adult and fawn mortality, no drought or inbreeding depression, and large initial population size, has an annual growth rate of nearly 9% with no risk of extinction in 100 years (Table 1, Fig. 1). At the other extreme, the most pessimistic scenario yields a population that is certain to be extinct within 70 years, with an annual growth rate of -13%. It is likely that the population currently is intermediate between these two extremes, with a considerable risk of extinction within the next 100 years. The models indicate that both fawn and adult mortality appear to be the primary factors influencing the viability of the pronghorn population in the Vizcaino Desert. In fact, the results suggest that the population is roughly equally sensitive to increases in both fawn and adult 22 Peninsular Pronghorn PHVA Report mortality. Nevertheless, the modelling efforts clearly show that reduced fawn mortality, even with high adult mortality, can lead to population growth or stability, while higher fawn mortality always results in population decline and instability (Table 1). Together, these data suggest that management actions should be directed specifically toward reducing pronghorn mortality, perhaps with an emphasis on fawns. It appears that this can be accomplished through predator control and/or semi-captive rearing. Detailed cost-benefit analyses must be undertaken to determine the most appropriate course of action. Furthermore, the deleterious consequences of inbreeding can be considerable, particularly when pronghorn populations are quite small (i.e., <1 00), and these effects must be recognized and appreciated by those responsible for subspecies management. Direct efforts should be made to establish and maintain larger wild population sizes in order to reduce the impact of inbreeding depression. Peninsular Pronghorn PHVA Report 23 Sample VORTEX Input File PRONG320.0UT ***Output Filename*** Y ***Graphing Files?*** N ***Each Iteration?*** 100 ***Simulations*** 100 ***Years*** 10 ***Reporting Interval*** 1 ***Populations*** Y ***Inbreeding Depression?*** H ***Model of Inbreeding Depression*** 3.140000 ***Number of Lethal Equivalents?*** Y ***EV correlation?*** 1 ***Types Of Catastrophes*** P ***Monogamous, Polygynous, or Hermaphroditic*** 2 ***Female Breeding Age*** 1 ***Male Breeding Age*** 10 ***Maximum Age*** 0.500000 ***Sex Ratio*** 2 ***Maximum Litter Size*** N ***Density Dependent Breeding?*** 25.000000 ***Population 1: Percent Litter Size 0*** 7.500000 ***Population 1: Percent Litter Size 1*** 67.500000 ***Population 1: Percent Litter Size 2*** 2.500000 ***EV--Reproduction*** 60.000000 ***Female Mortality At Age 0*** 14.770979 ***EV--FemaleMortality*** 10.000000 ***Female Mortality At Age 1*** 2.500000 ***EV--FemaleMortality*** 20.000000 ***Adult Female Mortality*** 5.000000 ***EV--AdultFemaleMortality*** 60.000000 ***Male Mortality At Age 0*** 14.770979 ***EV--MaleMortality*** 25.000000 ***Adult Male Mortality*** 6.250000 ***EV--AdultMaleMortality*** 1.000000 ***Probability Of Catastrophe 1*** 1.000000 ***Severity--Reproduction*** 0.500000 ***Severity--Survival*** N ***All Males Breeders?*** Y ***Answer--A--Known?*** 75.000000 ***Percent Males In Breeding Pool*** Y ***Start At Stable Age Distribution?*** 100 ***Initial Population Size*** 500 ***K*** 0.000000 ***EV--K*** N ***Trend In K?*** N ***Harvest?*** N ***Supplement?*** Y ***AnotherSimulation?*** 24 Peninsular Pronghorn PHVA Report Table 1. Peninsular pronghorn population analysis: initial population size = 100, no catastrophes, no inbreeding depression. Mortality(%) Adult File# Fawn ~ d' %Males rd r, (SD) P(E) N 100(SD) HJOo T(E) 201 60 10 15 50 .090 .088(.114) 0.0 488 (26) 0.940 - 210 75 .090 .086 (.114) 0.0 489 (23) 0.943 - 202 15 20 50 .057 .052 (.129) 0.0 464 (50) 0.924 - 211 75 .057 .052 (.129) 0.0 459 (54) 0.927 - 203 20 25 50 .023 .016 (.151) 0.028 321 (148) 0.840 - 212 75 .023 .015 (.152) 0.026 309 (148) 0.843 - 204 67.5 10 15 50 .047 .041 (.133) 0.0 438 (84) 0.919 - 213 75 .047 .042 (.134) 0.0 438 (84) 0.926 - 205 15 20 50 .013 .002 (.164) 0.090 205 (162) 0.770 - 214 75 .013 .004 (.160) 0.074 214 (154) 0.809 - 206 20 25 50 -.022 -.043 (.217) 0.704 42 (52) 0.559 60 215 75 -.022 -.040 (.215) 0.662 41 (50) 0.573 62 207 75 10 15 50 -.003 -.023 (.194) 0.424 101 (118) 0.682 65 216 75 -.003 -.021 (.191) 0.394 101 (119) 0.700 62 208 15 20 50 -.039 -.066 (.234) 0.900 20 (19) 0.509 52 217 75 -.039 -.064 (.233) 0.900 28 (57) 0.529 52 209 20 25 50 -.076 -.110 (.266) 1.0 - - 36 218 75 -.076 -.110 (.267) 0.998 2(-) 0.375 36 %Males refers to the proportion of males available for breeding; rd is the deterministic growth rate calculated from the life table; r, (SD) is the stochastic growth (standard deviation) calculated from the simulations; P(E) is the probability of population extinction after 100 years; N 100 (SD) is the mean fmal population size (standard deviation) of those simulated populations surviving to 100 years; H100 is the mean population heterozygosity after 100 years; and T(E) is the mean time to population extinction for those scenarios in which at least 20% of the populations went extinct. Peninsular Pronghorn PHVA Report 25 Table 2. Peninsular pronghorn population analysis: initial population size = 100, no catastrophes, heterosis model of inbreeding depression (3 .14 lethal equivalents). Table heading definitions are as in Table 1. Mortality Adult f File# Fawn ~ d' %Males rct 8 (SD) P(E) Nw0(SD) Hwo T(E) 273 60 10 15 50 .090 .073 (.111) 0.0 481 (31) 0.940 - 282 75 .090 .074(.111) 0.0 483 (28) 0.944 - 274 15 20 50 .057 .037 (.128) 0.01 419(108) 0.916 - 283 75 .057 .040 (.126) 0.0 428 (77) 0.923 - 275 20 25 50 .023 -.018 (.167) 0.41 145 (140) 0.815 74 284 75 .023 -.023 (.175) 0.49 151 (148) 0.826 72 276 67.5 10 15 50 .047 .028 (.131) 0.02 377(131) 0.921 - 285 75 .047 .027 (.132) 0.02 373 (139) 0.909 - 277 15 20 50 .013 -.035 (.185) 0.63 93 (126) 0.771 70 286 75 .013 -.030 (.173) 0.54 69 (97) 0.784 67 278 20 25 50 -.022 -.075 (.226) 0.98 25 (27) 0.755 52 287 75 -.022 -.073 (.220) 0.98 7 (4) 0.676 52 279 75 10 15 50 -.003 -.046 (.197) 0.80 94 (106) 0.735 68 288 75 -.003 -.051 (.200) 0.79 39 (53) 0.776 63 280 15 20 50 -.039 -.088 (.230) 1.0 - - 44 289 75 -.039 -.088 (.230) 1.0 - - 44 281 20 25 50 -.076 -.123 (.258) 1.0 - - 32 290 75 -.076 -.131 (.266) 1.0 - - 30 26 Peninsular Pronghorn PHVA Report Table 3. Peninsular pronghorn population analysis: initial population size= 200, no catastrophes, no inbreeding depression. Table heading definitions are as in Table 1. Mortality Adult File# Fawn ~ d' %Males fct rs (SD) P(E) N100(SD) HIOo T(E) 219 60 10 15 50 .090 .086 (.113) 0.0 489 (23) 0.950 - 228 75 .090 .087 (.113) 0.0 490 (22) 0.954 - 220 15 20 50 .057 .052 (.128) 0.0 460 (54) 0.943 - 229 75 .057 .052 (.127) 0.0 462 (52) 0.945 - 221 20 25 50 .023 .016 (.145) 0.01 348 (130) 0.906 - 230 75 .023 .016 (.146) 0.004 340 (137) 0.905 - 222 67.5 10 15 50 .047 .040 (.132) 0.0 441 (79) 0.942 - 231 75 .047 .040 (.133) 0.0 438 (76) 0.945 - 223 15 20 50 .013 .005 (.153) 0.028 253 (146) 0.867 - 232 75 .013 .005 (.154) 0.032 268 (155) 0.875 - 224 20 25 50 -.022 -.040 (.204) 0.554 54 (68) 0.611 68 233 75 -.022 -.041 (.206) 0.570 59(81) 0.670 69 225 75 10 15 50 -.003 -.020 (.183) 0.238 108 (123) 0.752 72 234 75 -.003 -.020 (.181) 0.260 120 (125) 0.775 71 226 15 20 50 -.039 -.065 (.224) 0.858 39 (60) 0.579 60 235 75 -.039 -.065 (.224) 0.856 30 (45) 0.583 61 227 20 25 50 -.076 -.107 (.258) 0.994 11 (4) 0.272 43 236 75 -.076 -.108 (.258) 0.990 13 (9) 0.266 42 Peninsular Pronghorn PHVA Report 27 Table 4. Peninsular pronghorn population analysis: initial population size = 200, no catastrophes, heterosis model of inbreeding depression (3 .14 lethal equivalents). Table heading definitions are as in Table 1. Mortality Adult File# Fawn ~ d' %Males rd r, (SD) P(E) Nwo(SD) Hwo T(E) 291 60 10 15 50 .090 .079 (.Ill) 0.0 489 (24) 0.951 - 300 75 .090 .078 (.113) 0.0 483 (27) 0.953 - 292 15 20 50 .057 .043 (.125) 0.0 437 (80) 0.944 - 301 75 .057 .043 (.125) 0.0 451 (81) 0.942 - 293 20 25 50 .023 -.004 (.154) 0.13 201 (145) 0.863 - 302 75 .023 -.004 (.150) 0.11 206 (147) 0.878 - 294 67.5 10 15 50 .047 .029 (.132) 0.01 382 (114) 0.936 - 303 75 .047 .033 (.130) 0.0 396 (102) 0.945 - 295 15 20 50 .013 -.018 (.163) 0.28 154 (139) 0.852 81 304 75 .013 -.018 (.162) 0.27 133 (131) 0.856 74 296 20 25 50 -.022 -.068 (.207) 0.95 26 (27) 0.777 62 305 75 -.022 -.067 (.211) 0.94 44 (61) 0.781 64 297 75 10 15 50 -.003 -.039 (.185) 0.56 75 (104) 0.758 70 306 75 -.003 -.042 (.186) 0.58 65 (104) 0.777 68 298 15 20 50 -.039 -.079 (.223) 0.96 27 (40) 0.665 57 307 75 -.039 -.086 (.225) 0.98 7 (0) 0.714 53 299 20 25 50 -.076 -.121 (.253) 1.0 - - 38 308 75 -.076 -.118 (.248) 1.0 - - 40 28 Peninsular Pronghorn PHVA Report Table 5. Peninsular pronghorn population analysis: initial population size = 100, catastrophe (1 00-year drought), no inbreeding depression. Table heading definitions are as in Table 1. Mortality Adult File# Fawn ~ d' %Males rd r, (SD) P(E) N100(SD) Hwo T(E) 237 60 10 15 50 .085 .079 (.134) 0.0 477 (50) 0.937 - 246 75 .085 .080 (.133) 0.0 475 (56) 0.939 - 238 15 20 50 .052 .046 (.146) 0.002 440 (89) 0.913 - 247 75 .052 .046 (.148) 0.002 437 (91) 0.916 - 239 20 25 50 .018 .006 (.173) 0.106 255 (164) 0.798 - 248 75 .018 .007 (.174) 0.078 254 (155) 0.798 - 240 67.5 10 15 50 .042 .034 (.151) 0.008 391 (120) 0.902 - 249 75 .042 .032 (.153) 0.014 395 (121) 0.904 - 241 15 20 50 .008 -.005 (.183) 0.178 171 (147) 0.760 - 250 75 .008 -.006 (.182) 0.182 166 (149) 0.762 - 242 20 25 50 -.027 -.054 (.236) 0.842 33 (35) 0.498 58 251 75 -.027 -.052 (.233) 0.808 39 (40) 0.557 56 243 75 10 15 50 -.008 -.030 (.207) 0.518 74 (94) 0.668 65 252 75 -.008 -.029 (.210) 0.498 69 (87) 0.665 62 244 15 20 50 -.044 -.072 (.248) 0.946 23 (28) 0.400 50 253 75 -.044 -.076 (.251) 0.950 19 (18) 0.488 48 245 20 25 50 -.081 -.123 (.280) 1.0 - - 32 254 75 -.081 -.120 (.281) 0.998 9(-) 0.346 33 Peninsular Pronghorn PHVA Report 29 Table 6. Peninsular pronghorn population analysis: initial population size = 100, catastrophe (100-year drought), heterosis model of inbreeding depression (3.14lethal equivalents). Table heading definitions are as in Table 1. Mortality Adult File# Fawn ~ 0' %Males rd r, (SD) P(E) N 100(SD) HJOo T(E) 309 60 10 15 50 .085 .066 (.132) 0.0 446 (89) 0.934 - 318 75 .085 .068 (.132) 0.0 469 (64) 0.938 - 310 15 20 50 .052 .022 (.148) 0.06 348 (142) 0.897 - 319 75 .052 .025 (.148) 0.05 357 (140) 0.908 - 311 20 25 50 .018 -.034 (.196) 0.62 90 (94) 0.809 67 320 75 .018 -.031 (.192) 0.56 99 (134) 0.783 67 312 67.5 10 15 50 .042 .014 (.154) 0.07 293 (165) 0.877 - 321 75 .042 .016 (.157) 0.08 301 (166) 0.878 - 313 15 20 50 .008 -.041 (.199) 0.69 57 (92) 0.732 65 322 75 .008 -.037 (.196) 0.71 89 (90) 0.785 71 314 20 25 50 -.027 -.085 (.241) 1.0 - - 46 323 75 -.027 -.084 (.243) 0.99 3 (-) 0.0 47 315 75 10 15 50 -.008 -.051 (.208) 0.83 43 (64) 0.726 62 324 75 -.008 -.056 (.210) 0.86 36 (42) 0.716 61 316 15 20 50 -.044 -.099 (.249) 1.0 - - 40 325 75 -.044 -.099 (.244) 1.0 - - 40 317 20 25 50 -.081 -.133 (.274) 1.0 - - 30 326 75 -.081 -.142 (.278) 1.0 - - 28 30 Peninsular Pronghorn PHVA Report Table 7. Peninsular pronghorn population analysis: initial population size= 200, catastrophe (1 00-year drought), no inbreeding depression. Table heading definitions are as in Table 1. Mortality Adult File# Fawn !? rJ' %Males rd r, (SD) P(E) N 100(SD) Hwo T(E) 255 60 10 15 50 .085 .079 (.133) 0.0 478 (50) 0.949 - 264 75 .085 .079 (.134) 0.0 473 (55) 0.951 - 256 15 20 50 .052 .046 (.144) 0.0 445 (78) 0.937 - 265 75 .052 .044 (.146) 0.0 438 (94) 0.937 - 257 20 25 50 .018 .011 (.162) 0.022 300 (145) 0.881 - 266 75 .018 .008 (.166) 0.030 284 (151) 0.869 - 258 67.5 10 15 50 .042 .034 (.152) 0.002 393 (117) 0.932 - 267 75 .042 .035 (.150) 0.0 415 (104) 0.939 - 259 15 20 50 .008 -.004 (.174) 0.092 196(156) 0.831 - 268 75 .008 -.005 (.176) 0.106 202 (157) 0.828 - 260 20 25 50 -.027 -.048 (.218) 0.692 45 (54) 0.616 66 269 75 -.027 -.049 (.221) 0.682 37 (43) 0.633 66 261 75 10 15 50 -.008 -.028 (.200) 0.358 86 (112) 0.721 70 270 75 -.008 -.030 (.203) 0.402 85 (107) 0.730 69 262 15 20 50 -.044 -.069 (.239) 0.890 34 (34) 0.639 58 271 75 -.044 -.070 (.238) 0.898 25 (40) 0.521 58 263 20 25 50 -.081 -.114 (.268) 0.996 14 (3) 0.546 40 272 75 -.081 -.111 (.269) 0.996 5 (1) 0.0 42 Peninsular Pronghorn PHVA Report 31 Table 8. Peninsular pronghorn population analysis: initial population size = 200, catastrophe (1 00-year drought), heterosis model of inbreeding depression (3 .14 lethal equivalents). Table heading definitions are as in Table 1. Mortality Adult File# Fawn !? d' %Males rct r, (SD) P(E) N100(SD) HIOo T(E) 327 60 10 15 50 .085 .071 (.132) 0.0 472 (57) 0.949 - 336 75 .085 .069 (.133) 0.0 460 (71) 0.949 - 328 15 20 50 .052 .031 (.147) 0.02 395 (127) 0.928 - 337 75 .052 .033 (.143) 0.0 402 (127) 0.929 - 329 20 25 50 .018 -.020 (.179) 0.32 137 (134) 0.844 78 338 75 .018 -.013 (.172) 0.20 154 (133) 0.860 79 330 67.5 10 15 50 .042 .024 (.148) 0.0 355 (142) 0.931 - 339 75 .042 .024 (.149) 0.01 349 (144) 0.930 - 331 15 20 50 .008 -.032 (.190) 0.49 109 (108) 0.831 73 340 75 .008 -.025 (.186) 0.41 150 (156) .843 77 332 20 25 50 -.027 -.080 (.227) 0.99 12 (-) 0.722 58 341 75 -.027 -.077 (.231) 0.96 12 (5) 0.514 59 333 75 10 15 50 -.008 -.055 (.212) 0.77 53 (62) 0.725 67 342 75 -.008 -.048 (.205) 0.72 78 (107) 0.821 68 334 15 20 50 -.044 -.099 (.240) 1.0 - - 47 343 75 -.044 -.092 (.235) 0.99 22(-) 0.871 49 335 20 25 50 -.081 -.137 (.259) 1.0 - - 34 344 75 -.081 -.133 (.268) 1.0 - - 36 32 Peninsular Pronghorn PHVA Report Figure Legends Figure 1. Probability of population extinction (a) and final population size (b) for pronghorn populations with juvenile mortality = 60% and an initial population size of 100 under different adult mortality scenarios: A: 10% female, 15% male B: 15% female, 20% male C: 20% female, 25 %male (A-C)/F: Scenarios incorporating inbreeding depression. In these and all subsequent figures, only scenarios incorporating 50% of adult males into the breeding pool are shown. Figure 2. Probability of population extinction (a) and final population size (b) for pronghorn populations with juvenile mortality = 60% and an initial population size of 200 under different adult mortality scenarios. Curves defined as in Figure 1. Figure 3. Probability of population extinction (a) and final population size (b) for pronghorn populations with juvenile mortality= 67.5% and an initial population size of 100 under different adult mortality scenarios. Curves defined as in Figure 1. Figure 4. Probability of population extinction (a) and final population size (b) for pronghorn populations with juvenile mortality= 67.5% and an initial population size of200 under different adult mortality scenarios. Curves defined as in Figure 1. Figure 5. Probability of population extinction (a) and final population size (b) for pronghorn populations with juvenile mortality= 75% and an initial population size of 100 under different adult mortality scenarios. Curves defined as in Figure 1. Figure 6. Probability of population extinction (a) and final population size (b) for pronghorn populations with juvenile mortality= 75% and an initial population size of200 under different adult mortality scenarios. Curves defined as in Figure 1. Figure 7. Probability of population extinction (a) and final population size (b) for pronghorn populations with juvenile mortality = 60%, an initial population size of 100, and exposure to drought under different adult mortality scenarios. Curves defined as in Figure 1. Figure 8. Probability of population extinction (a) and final population size (b) for pronghorn populations with juvenile mortality= 60%, an initial population size of200, and exposure to drought under different adult mortality scenarios. Curves defined as in Figure 1. Figure 9. Probability of population extinction (a) and final population size (b) for pronghorn populations with juvenile mortality= 67.5%, an initial population size of 100, and exposure to drought under different adult mortality scenarios. Curves defined as in Figure 1. Peninsular Pronghorn PHVA Report 33 Figure 10. Probability of population extinction (a) and final population size (b) for pronghorn populations with juvenile mortality= 67.5%, an initial population size of200, and exposure to drought under different adult mortality scenarios. Curves defined as in Figure 1. Figure 11. Probability of population extinction (a) and final population size (b) for pronghorn populations with juvenile mortality= 75%, an initial population size of 100, and exposure to drought under different adult mortality scenarios. Curves defined as in Figure 1. Figure 12. Probability of population extinction (a) and final population size (b) for pronghorn populations with juvenile mortality= 75%, an initial population size of200, and exposure to drought under different adult mortality scenarios. Curves defined as in Figure 1. 34 Peninsular Pronghorn PHVA Report Fig. 1. Juvenile Mortality= 60% N0 = 100 1.o (a) 500 (b) 0.9 Adult Mortality: 450 --A c 0.8 --B 400 0 -A- c 0.7 -o- AIF 350 uc Q) :;::; -o- B/F .!::! X 0.6 ----b- C/F Cl) 300 w c '+- 0 0 0.5 :;::; 250 >. ~ :t: :::l 0.4 c.. 200 :cro 0 ..c 0.3 11. 150 ,_0 11. 0.2 100 0.1 50 0.0 0 0 20 40 60 80 100 0 20 40 60 80 100 Year Year Fig. 2. Juvenile Mortality= 60% N0 = 200 1.o (a) 500 (b) 0.9 450 c 0.8 400 0 0.7 350 uc Q) :;::; N X 0.6 i:75 300 w c '+- 0 0 0.5 :;::; 250 >. ro :t: :c 0.4 :::lc.. 200 ro 0 ..c 0.3 11. 150 ,_0 11. 0.2 100 0.1 50 0.0 0 0 20 40 60 80 100 0 20 40 60 80 100 Year Year Peninsular Pronghorn PHVA Report 35 Fig. 3. Juvenile Mortality= 67.5% N0 = 100 1.o (a) 500 (b) 0.9 Adult Mortality: 450 --A c: 0.8 --B 400 0 -6- c :;:::; (.) 0.7 -a- A/F 350 c (]) :;:::; -a- B/F N X 0.6 -&- C/F i:i5 300 UJ c 0 0 0.5 :;:::; 250 - ro >. :!:::: :J 200 0.4 0. :.aro 0 .0 0.3 0.. 150 ....0 0.. 0.2 100 0.1 50 20 40 60 80 100 20 40 60 80 100 Year Year Fig. 4. Juvenile Mortality= 67.5% N0 = 200 1.o (a) 500 (b) 0.9 450 c 0.8 400 0 :;:::; (.) 0.7 350 c (]) :;:::; N X 0.6 i:i5 300 UJ c 0 0 0.5 :;:::; 250 - ro ~ 0.4 :J 200 :.a 0. ro 0 .0e 0.3 0.. 150 0.. 0.2 100 0.1 50 0.0 0 20 40 60 80 100 20 40 60 80 100 Year Year 36 Peninsular Pronghorn PHVA Report Fig. 5. Juvenile Mortality= 75% N0 = 100 1.o (a) 500 (b) 0.9 Adult Mortality: 450 --A c:: 0.8 --s 400 0 ~c :;::::; -o-- A/F u 0.7 Q} 350 c:: -o-- BIF :;::::; N >< 0.6 --t!r- CIF Ci5 300 w c:: 0 0 0.5 :;::::; 250 ->. ro :t= 0.4 ::I 200 :c 0.. ro 0 .c 0.3 0.. 150 .....0 0.. 0.2 100 0.1 50 0.0 0 20 40 60 80 100 20 40 60 80 100 Year Year Fig. 6. Juvenile Mortality= 75% N0 = 200 (a) (b) 1.0 500r-~--r-~--r-~--r-~--r-~--~ 0.9 450 c:: 0.8 400 0 :;::::; u 0.7 350 c:: Q} :;::::; N >< 0.6 Ci5 300 w c:: 0 0 0.5 :;::::; 250 - ro ~ 0.4 ::I 200 :0 0.. ro 0 .c 0.3 0.. 150 .....0 0.. 0.2 100 0.1 50 0.0 0 20 40 60 80 100 20 40 60 80 Year Year Peninsular Pronghorn PHVA Report 37 Fig. 7. Juvenile Mortality= 60% N0 = 100; Drought 1.o (b) (a) 500 0.9 Adult Mortality: 450 ------A c: 0.8 --s 400 0 --4-- c 0.7 -o- A/F 350 uc: (]) :;:::; -o- B/F .!:::! >< 0.6 --t:!r- C/F (f) 300 w c: 0 -0 0.5 :;:::; 250 :!::>- ~ 0.4 ::::l 200 0. :0 0 ro (L ..0e 0.3 150 (L 0.2 100 0.1 50 0.0 0 0 20 40 60 80 100 0 20 40 60 80 100 Year Year Fig. 8. Juvenile Mortality= 60% N0 = 200; Drought (a) (b) 1.0 500 0.9 450 c: 0.8 400 0 :;:::; () 0.7 350 c: (]) :;:::; N >< 0.6 U5 300 w c: 0 0 0.5 :;:::; 250 ->- ro ;'!:::! 0.4 ::::l 200 :0 0. ro 0 ..0 0.3 (L 150 ....0 (L 0.2 100 0.1 50 0.0 0 0 20 40 60 80 100 0 20 40 60 80 100 Year Year 38 Peninsular Pronghorn PHVA Report Fig. 9. Juvenile Mortality= 67.5% N0 = 100; Drought (a) (b) 1.0 500 0.9 Adult Mortality: 450 --A c: 0.8 --B 400 0 ---A- c :;:::. (.) 0.7 -o- A/F 350 c: Q) :;:::. --o- B/F N X 0.6 -e.- C/F U5 300 w c: 0 0 0.5 :;:::. 250 - ro ~ 0.4 :::J 200 :0 0. ro 0 ..0 0.3 c.. 150 ,_0 c.. 0.2 100 0.1 50 0.0 0 20 40 60 80 100 20 40 60 80 100 Year Year Fig. 10. Juvenile Mortality= 67.5% N0 = 200; Drought (a) (b) 1.0 500 0.9 450 c: 0.8 400 0 :;:::. (.) 0.7 350 c: Q) :;:::. N X 0.6 U5 300 w c: 0 0 0.5 :;:::. 250 ->. jQ :!:::! 0.4 :::J 200 :0 0. ro 0 ..0 0.3 c.. 150 ,_0 c.. 0.2 100 0.1 50 0.0 0 20 40 60 80 100 20 40 60 80 100 Year Year Peninsular Pronghorn PHVA Report 39 Fig. 11. Juvenile Mortality= 75% N0 = 100; Drought (a) (b) 1.0 500 Adult Mortality 0.9 --fl- A 450 --fl-8 c: 0.8 -A- c 400 0 :;::::l -o- A/F (.) 0.7 -o-- 8/F Q) 350 c: N :;::::l -6-- C/F X 0.6 U5 300 w c: 0 0 0.5 :;::::l 250 ->- ro :!::: 0.4 ::l 200 :0 c.. co 0 .c 0.3 !1. 150 .._0 !1. 0.2 100 0.1 50 0.0 0 20 40 60 80 100 20 40 60 80 100 Year Year Fig. 12. Juvenile Mortality= 75% N0 = 200; Drought (a) 1.0 500 (b) 0.9 450 c: 0.8 400 0 0.7 350 tsc: Q) :;::::l .!::! X 0.6 (f) 300 w c: 0 0 0.5 :;::::l 250 - ~ ~ 0.4 ::l 200 :0 c.. co 0 .c 0.3 !1. 150 .._0 !1. 0.2 100 0.1 50 0.0 0 20 40 60 80 100 20 40 60 80 100 Year Year 40 Peninsular Pronghorn PHVA Report EVALUACION DE VIABILIDAD DE LA POBLACION Y DEL HABITAT DEL BERRENDO PENINSULAR (Antilocapra americana penins ularis) POPULATION AND HABITAT VIABILITY ASSESSMENT FOR THE PENINSULAR PRONGHORN (Antilocapra americana peninsularis) La Paz, Baja California Sur, Mexico Centro de Investigaciones Biologicas del Noroeste, S.C. 16-18 de Noviembre de 1994 SECCION3 DISTRIBUCION Y ESTADO DE CONSERVACION DISTRIBUTION AND CONSERVATION STATUS DISTRIBUCION Y ESTADO DE CONSERVACION Distribucion Historica La subespecie endemica de la peninsula (Hally Kelson 1959) se distribuia historicamente en gran parte de ella, siendo los registros marginales los siguientes: Cafion de la Esperanza (Elliot 1903); Bahia de San Felipe (Elliot 1903); Tierra adentro de la Bahia de Santa Rosalia (Townsend 1912); al sur, mas alla de la Bahia de Magdalena (Nelson 1925), 73 km al sur de Calmalli; Bahia de Vizcaino (Nelson 1925), Desierto de Vizcaino (Orihuela 1977); San Quintin (Elliot 1903). Se 2 estima que historicamente se distribuyo en aproximadamente 40,000 km • El impacto de las actividades humanas como son la construccion de la carretera transpeninsular, exploraciones de Pemex en Vizcaino, construccion del acueducto Vizcaino Bahia Tortugas y de caminos vecinales en la region, entre 1950 y 1980, fueron factores que determinaron la rapida declinacion de las poblaciones del berrendo. Aunque la distribucion del berrendo peninsular presentada por Hall (1981) no incluye el sur de California, existe la posibilidad de que esta subespecie en realidad ocupara tambien esta region. Distribucion Actual La distribucion actual de acuerdo a los registros realizados entre 1977 y 1994, es de 362,385 ha aproximadamente (SEDESOL, 1993, mas obs pers. B. Sanabria). Dicha area cubre una superficie plana, topograficamente simple, incluida entre el paralelo 28 en el norte, 113 grados y 18 minutos en el este, 26 grados 47 minutos en el sur, y el meridiana 114 grados 30 minutos en el oeste (Figura 3-1, pagina 45). Lineas de Investigacion - Es necesario desarrollar estudios de investigacion que establezcan la identidad taxonomica en las poblaciones distribuidas en la frontera entre Baja California Norte y California, E. U. Entre las herramientas a usar puede estar la nueva tecnica para extraer DNA de cavidad pulpar de dientes de ejemplares de museo. - Es necesario identificar las areas geognificas de crianza y distribucion estacional, a traves de observaciones directas y de tecnicas de radiotelemetria. - Dada la variacion en el esfuerzo de muestreo en los censos en el pasado, es necesario estandarizar los metodos y periodos en que se realizan los censos y hacer un analisis mas profunda de los datos ya existentes. En los censos se podrian utilizar metodos como los instrumentados en el caso de Sonora, que incluyen dobles conteos, transectos lineales, radiotelemetria, correccion para visibilidad (visibility bias). Peninsular Pronghorn PHVA Report 43 Estado de Conservacion del Berrendo El berrendo peninsular hist6ricamente fue protegido legalmente por primera vez en 1922, afio en el cual el Presidente Obregon decret6 su veda permanente. Posteriormente, en 1951, se decret6 la Ley Federal de Caza, la cualle di6 continuidad a aquella disposici6n. Despues, a nivel intemacional esta subespecie fue incluida en el Red Data Book de la IUCN en 1988. En el "Federal Register" el USFWS incluy6 igualmente al berrendo peninsular bajo estado de "amenazado". Asi mismo, en el Diario Oficial de 1991 se reiter6 la protecci6n legal de esta subespecie. El mas reciente decreta oficial en nuestro pais en el cual se le otorga protecci6n es el del21 de mayo de 1994. El impacto de estas disposiciones legales ha permitido al menos la sobrevivencia de una pequeiia poblaci6n estimada en aproximadamente 200 ejemplares hasta la fecha. Se han recopilado datos en 23 censos en el area de distribuci6n actual en un periodo que va desde 1977 a 1986 (Figura 3-2, pagina 46). A partir de esta fecha hubo un periodo de cuatro aiios en los que se carece de datos, siendo hasta el afio de 1990 en el que nuevamente se continuaron a partir del mes de noviembre, realizandose seis censos mas. De estos, el mas notable fue el de noviembre de 1993. En este censo se observaron 175 animales: 48 machos, 56 hembras y 71 no identificados. En el grafica anterior, mostramos el censo estandorizado para una observaci6n anual. Aunque la tendencia parece indicar un comportamiento constante de la poblacion con algunos momentos de gran abundancia como 1985, 1986, y 1993. Esto puede ser resultado de esfuerzos de muestreo desiguales, distintas areas y extensiones muestreadas, y diferentes capacidades de detecci6n por los observadores, o bien representar realmente el comportamiento de la poblaci6n durante ese lapso. 44 Peninsular Pronghorn PHVA Report U.S. A. ~ Present distribution. Former distribution. ~ Biosphere Reserve boundary. SCALE 0 25 50 Km Figure 3-1. Historic and present distribution of the peninsular pronghorn in the Lower California peninsula, Mexico. Peninsular Pronghorn PHVA Report 45 200 180 160 140 Q) .~ (j) 120 c 0 :;::::; 100 ro ::Jc.. 80 0 0.. 60 40 20 0 L <') <') <') ~ N co 0 0 N (0 co ~ co N N ' Figure 3-2. Population census of the peninsular pronghorn in the Vizcaino Desert from 1977 to 1994. 46 Peninsular Pronghorn PHVA Report DISTRIBUTION AND CONSERVATION STATUS Historical Distribution The peninsular pronghorn (Hall and Kelson 1959) was historically distributed over a large part of the peninsula, based on the following records: Canon de la Esperanza (Elliott 1903); Bahia de San Felipe (ibid.); Litoral de la Bahia de Santa Rosalia (Townsend, 1912); to the south beyond Bahia de Magdelena (Nelson, 1925), 73 km to the south ofCalmalli; Bahia de Vizcaino (Nelson, 1925); Desierto de Vizcaino; San Quentin (Elliot, 1903). It has been estimated that the subspecies was distributed over approximately 40,000 km2. The impact of human activities, such as construction of the transpeninsular highway, explorations by PEMEX in the Vizcaino area, and construction of the aqueduct Vizcaino-Bahia Tortugas and small roads in the region, were factors contributing to the rapid decline of pronghorn populations during 1950-1970. Although the distribution of the peninsular pronghorn presented by Hall (1981) does not include southern California, there is a possibility that this subspecies was also distributed in this area. Current Distribution The present occupied range is approximately 362,385 ha (SEDESOL 1993; B. Sanabria pers. obs.). This area covers a topographically flat expanse, including the 28th parallel to the north, 113 degrees, 18 minutes East, 26 degrees, 4 7 minutes to the south. Towards the west it is bounded by the meridian at 114 degrees, 30 minutes (Figure 3-1, page 45). Research Needs • Develop research studies which establish the taxonomic identity of those populations distributed in the border between Baja California Norte and California, USA. Among the techniques that can be used for this study, a new method involves extraction of DNA from tooth pulp of museum specimens. • Identify the geographical areas of rearing and the seasonal distribution, through the direct observations of females isolated from the group and through radiotelemetry. • Due to the variation in the survey efforts in the past, it is necessary to standardize the methods and seasons in which the censuses are done and to conduct more thorough analysis of the data which already exist. In the censuses, it may be possible to use methods similar to those implemented in the case of the Sonoran pronghorn, which include double counts, lineal transects, rediotelemetry, and visibility bias. Peninsular Pronghorn PHVA Report 47 Conservation Status of the Pronghorn Historically, the peninsular pronghorn was protected by law for the first time in 1922 when President Obregon decreed that hunting was prohibited. In 1951, Ley Federal de Caza, or the Federal Hunting Law, continued this level of protection. At the international level, the subspecies was listed in the Convention on International Trade in Endangered Species (CITES) of wild flora and fauna in 1975 as an Appendix I species and included in the IUCN Red Data Book in 1988. Appendix I species are those threatened with extinction. Similarly, in the Diario Oficial of 1991, its level of protection was decreed and the last official decree for protection of the pronghorn was 21 May, 1994. These legal actions have permitted the survival of a small population, estimated at 100-200 animals. Data from 23 censuses (1977 to 1986) in the occupied range have been compiled (Figure 3-2, page 46). There was a period 1986-1990 in which there was no census. Censuses were resumed in November 1990 and six censuses have been conducted since then. Among these six censuses, the most important was the one done in November 1993. In this census, 175 animals were observed: 48 males, 56 females, and 71 unidentified. In the preceding graph, we show the census standardized for one annual observation. The population appears stable, with some increases in 1985, 1986, and 1993. This can be the result of different survey efforts, different areas and extents of coverage, and different capacities of detection for the observers, as well as real population trends during this period. 48 Peninsular Pronghorn PHVA Report EVALUACION DE VIABILIDAD DE LA POBLACION Y DEL HABITAT DEL BERRENDO PENINSULAR (Antilocapra americana penins ularis) POPULATION AND HABITAT VIABILITY ASSESSMENT FOR THE PENINSULAR PRONGHORN (Antilocapra americana peninsularis) La Paz, Baja California Sur, Mexico Centro de Investigaciones Biologicas del Noroeste, S.C. 16-18 de Noviembre de 1994 SECCION 4 CALIDAD DE HABITAT Y TENDENCIAS HABITAT QUALITY AND TRENDS CALIDAD DE HABITAT Y TENDENCIAS Antecedentes Desde la decada de los 60's el area de distribucion del berrendo peninsular ha tenido variaciones, 2 en un principio se estimo que ocupada 8000 km • En la actualidad se estima que entre el area ocupada y la que potencialmente pueden usar los berrendos, se alcanza una extension de 5000 2 km , ya que en su mayor parte conserva las condiciones de vegetacion, topografia, y accesibilidad para esta subespecie. Existen estudios que presentan informacion sobre la dieta del berrendo. El primero (Cancino 1988) ofrece un listado de especies vegetales consumidas. El segundo (Cancino in press) describe los habitos alimentarios basados en analisis de excretas. Referenda general sobre las causas por las que se modifico la estructura y extension del habitat en Vizcaino (Jaramillo 1989). Otra publicacion (Cancino 1988) que cuantifica de los cambios en la extension del habitat historico y causas asociadas al hombre. Necesidades de Investigacion 1) Se necesita informacion sobre los cambios en el uso del suelo y desarrollos urbanos para finalmente relacionarlo con la distribucion historica y actual. Se require de un estudio cuantitativo de la transformacion y/o reduccion del habitat debido a los siguientes factores. -Agricultura -Nuevos centros de poblacion -Inundacion de salitrales (industriales) -Construccion de acueductos y diques -Construccion de caminos y brechas -Desarrollos turisticos -Uso de recursos que afectan la calidad del habitat: lefia y posterias; mineria y extractivas; uso de agua, y sobrexplotacion de acuiferos. Herramientas: Mapas INEGI, DETENAL, SARH, COTECOCA, Reforma Agraria (tenencia de la tierra), usos del suelo y planes de desarrollo urbano (gobiemo del estado ), Pemex. 2) Informacion sobre datos climaticos lo mas antiguos posibles. Peninsular Pronghorn PHVA Report 51 3) Determinar caracteristicas estructurales de la vegetaci6n actuales para conocer los requerimientos del habitat cuantitativamente (nfunero de estratos, especies dominantes, y fenologia) 4) Determinar las caracteristicas topognificas y de altitud donde ha sido registrado el berrendo. 5) Determinar la selecci6n de sitios de pernoctaci6n y reproducci6n. 6) Capacidad de carga. 7) Impacto de la fauna asociada: posibles depradadores (coyote, aguila, zorra) y competidores (liebres y ganado) Esta informacion podra ampliar el conocimiento de la disponibilidad del ha. bitat adecuado actual y del hist6rico, y las causas posibles del decremento de la poblaci6n, y ayudani a determinar las posibilidades de reintroducci6n en estas areas. 52 Peninsular Pronghorn PHVA Report HABITAT QUALITY AND TRENDS Background Since the 1960's the distribution area of the peninsular pronghorn has been variable, but was 2 2 generally considered to be about 8000 km • At present it is estimated that about 5000 km of pronghorn habitat exists, based on vegetation conditions, topography, and accessibility for this subspecies. Only a portion of this available habitat is currently used by pronghorn. Two studies present information about the diet of pronghorn. The first is a listing of plant species consumed (Cancino 1988). The second describes food habits based on fecal analysis (Cancino in press). Jaramillo (1989) discusses the reasons why the size and shape of habitat in the Vizcaino Desert was modified. Furthermore, Cancino (1988) quantifies changes in size and extent of the historical habitat and human related factors causing the changes. Research Needs 1. We need information about changes in land use and urban development to relate them to the historical distribution of the pronghorn. We require quantitative study of the transformation and/or reduction of the habitat due to the following: - Agriculture -New human population centers -Mining of salt - Construction of aqueducts and canals - Construction of paved and unpaved roads - Development of tourism -Use of resources which affect habitat quality: firewood, power lines; mining; extractive industries; use of water; overexploitation of aquifers. Tools: Maps ofNational Institute of Geography, DETENAL, SARH, COTECOCA, Agrarian Reform (land tenancy), land use, and urban development plan (state government), Petroleos Mexicanos (PEMEX) 2. Historical information about climatological data. 3. To determine the structural characteristics of the vegetation to understand the quantitative habitat requirements of the pronghorn (habitat strata, dominant species, and phenology) 4. To determine topographic and elevation characteristics where the pronghorn has been recorded. Peninsular Pronghorn PHVA Report 53 5. To determine the habitat characteristics of selected sites. 6. Carrying capacity. 7. Impact of associated wildlife like potential predators (coyote, eagle, and fox) and competitors (hares and cattle). This information should increase our knowledge of the distribution and characteristics of the present and historical habitat, identify possible causes of the population decline, and help determine the potential for reintroduction into unoccupied habitat. 54 Peninsular Pronghorn PHVA Report EVALUACION DE VIABILIDAD DE LA POBLACION Y DEL HABITAT DEL BERRENDO PENINSULAR (Antilocapra americana peninsularis) POPULATION AND HABITAT VIABILITY ASSESSMENT FOR THE PENINSULAR PRONGHORN (Antilocapra americana peninsularis) La Paz, Baja California Sur, Mexico Centro de Investigaciones Biologicas del Noroeste, S.C. 16-18 de Noviembre de 1994 SECCION 5 AMENAZAS THREATS AMENAZAS Entre las amenazas ala sobrevivencia del berrendo en el Desierto del Vizcaino que se han observado en las ultimas decadas yen la actualidad, estan las siguientes: A. Transformaci6n y reducci6n de habitat B. Caceria ilegal C. Ganaderia extensiva D. Factores naturales Transformacion y Reduccion de Habitat Dentro de las actividades humanas que han contribuido a la transformaci6n y reducci6n del habitat del berrendo, se encuentran las siguientes: Zonas inundadas 33,000 ha Ganaderia extensiva 500,000 ha Agricultura 6,847 ha Brechas y caminos 2,200 ha Asentamientos humanos 1,583 ha Aislamiento de la Mesa de la Choya 25,550 ha Desde la decada de los 60's el area de distribuci6n del berrendo peninsular ha tenido variaciones, 2 en un principio se estim6 que ocupada 8000 km • En la actualidad se estima que entre el area ocupada y la que potencialmente pueden usar los berrendos, se alcanza una extension de 5000 2 km , ya que en su mayor parte conserva las condiciones de vegetaci6n, topografia, y accesibilidad para esta subespecie. Caceria Ilegal La causa mas evidente para la disminuci6n del berrendo en el Desierto del Vizcaino es la intensa y continua caceria ilegal a la que ha estado sometido, la anterior se infiere por las siguientes razones: relatos y denuncias de los habitantes del area sobre; incursiones de cazadores ilegales, infracciones por violaci6n a la legislaci6n vigente en materia de caceria, decomiso de armas a cazadores furtivos y restos de berrendos con evidencia de caza ilegal y multiples rastros (rodadas y casquillos do alto poder) en las zonas que actualmente habita el berrendo. Ganaderia Extensiva Existen aproximadamente 5400 cabezas de ganado vacuno en la porci6n oriental del Desierto del Vizcaino, dentro del area de distribuci6n del berrendo de las cuales mas del 60% del ganado se localiza en la principal zona de concentraci6n del berrendo. Tambien hay varias manadas de burros y caballos. El ganado domestico puede representar una fuerte competencia para el Peninsular Pronghorn PHVA Report 57 berrendo sobre todo en la epoca de secas en que ambos se concentran en los cauces de los arroyos y es probable que el ganado constituya ademas una fuente de transmisi6n de emfermedades para el berrendo a las que este no sea resistente. Factores Naturales Los factores naturales tambien pueden estar influyendo en la reducci6n de la poblaci6n relictual del berrendo, principalmente la escasa e irregular precipitaci6n pluvial y los prolongados periodos de sequias, los que disminuyen la tasa de reproducci6n y viabilidad de crias, si a esto aunamos la gran abundancia de coyotes, los que pueden causar, como se ha demostrado para otras subespecies de berrendo, una gran depredaci6n de crias recien nacidas, retardando de esta manera la recuperaci6n del berrendo peninsular. Por otra parte lo reducido y aislado de la poblaci6n e incluso de las manadas de berrendo pueden estar provocando la reproduccion entre parientes cercanos, con el consecuente incremento en la expresi6n de caracteres recesivos y por ende de malformaciones congenitas, lo cual puede ser una causa adicional de mortalidad. Conclusiones En base al conocimiento actual de la subespecie se considera que los principales amenazas a la sobrevivencia del berrendo en la actualidad se encuentran los siguientes: 1. Caceria 2. Modificaci6n de habitat 3. Mortalidad de crias Acciones Entre las acciones mas urgentes que deben de realizarse para la disminuci6n o control de las amenazas del berrendo peninsular se recomienda las siguientes: 1. Campanas permanentes de vigilancia, divulgaci6n y educaci6n para prevenir y controlar la caceria furtiva e impactos ambientales negatives, asi como concientizar y lograr la participaci6n activa de los habitantes locales en la recuperaci6n de la subespecie. 2. Gestiones para la protecci6n del habitat. Una medidad decisiva es la operaci6n de la Reserva de Biosfera el Vizcaino. 3. Llevar a cabo las gestiones para la coordinaci6n intersectorial, asi como para conseguir los fondos para la implementaci6n del plan. 4. En cuanto a investigaci6n se recomienda evaluar la competencia con el ganado domestico, la depredaci6n y causas de mortalidad de crias. 5. Evaluaci6n de la reducci6n del habitat en relaci6n ala poblaci6n del berrendo peninsular. 58 Peninsular Pronghorn PHVA Report THREATS Threats to survival of pronghorn observed in the last decade in the Vizcaino Desert are the following: A. Transformation and reduction of habitat B. Poaching C. Extensive livestock production D. Natural factors Transformation and Reduction of Habitat The following are human activities that have contributed to the transformation and reduction of pronghorn habitat: Zones of inundation 33,000 ha Livestock production 500,000 ha Agriculture 6,847 ha Paved and unpaved roads 2,200 ha Human settlements 1,583 ha Isolation of the Mesa de la Choya 25,550 ha Since the 1960's the distribution area of the peninsular pronghorn has been variable, but was 2 2 generally considered to be about 8000 km • At present it is estimated that about 5000 km of pronghorn habitat exists, based on vegetation conditions, topography, and accessibility for this subspecies. Only a portion of this available habitat is currently used by pronghorn. Poaching The most apparent cause for the decline of the pronghorn in the Desert is intensive and continuous illegal hunting. We believe this to be true based on the following evidence: reports and official complaints by local inhabitants about the presence and actions of illegal hunters, violations of hunting legislation, confiscation of illegal firearms, and finding the remains of pronghorns killed by poachers and several trails (vehicle tracks and empty shotgun shells) in current pronghorn habitat. Extensive Livestock Production Approximately 5400 livestock occur in the Vizcaino Desert within the distribution area of the pronghorn. More than 60% of these livestock and many groups of burros and horses are located in the principal concentration zone of the pronghorn. The livestock are strong competitors for the food of pronghorn, especially in the dry season when both concentrate along the arroyos. Peninsular Pronghorn PHVA Report 59 Livestock may also be a source of disease transmission to the pronghorn, disease to which they are not resistant. Natural Factors Natural factors, principally the low and irregular precipitation and prolonged dry periods, may also be reducing the remnant population of pronghorn by decreasing the reproductive rate and increasing mortality of the fawns. Another natural threat is the high population density of coyotes which may prey on recently born fawns, as has been demonstrated in studies of other pronghorn subspecies, thereby slowing or preventing population recovery of the peninsular pronghorn. The small and isolated populations of pronghorn may be resulting in inbreeding between closely related individuals, with consequent increase in recessive characters and congenital malformations, which may represent an additional cause of mortality. Conclusions Based on the current knowledge of the subspecies we consider that the following factors represent the main threats for the survival of the pronghorn: 1. Poaching 2. Habitat modification 3. Fawn mortality Actions Within the most urgent needs that have to be developed for the reduction or control of the peninsular pronghorn's threats, we recommend the following: 1. Permanent campaigns of patrolling, awareness and education to prevent and control the illegal hunting and the negative environmental impacts, and to increase consciousness and to make possible the active participation of the local inhabitants in recovery of the subspecies. 2. Actions to protect pronghorn habitat. One decisive action is operation of the Vizcaino Biosphere Reserve. 3. To coordinate actions between sectors, and to obtain funding to implement the plan. 4. In terms of research, it is recommended to evaluate the degree of competition with livestock, depredation and the main causes of fawn mortality. 5. Evaluation of reduction ofhabitat of the peninsular pronghorn population. 60 Peninsular Pronghorn PHVA Report EVALUACION DE VIABILIDAD DE LA POBLACION Y DEL HABITAT DEL BERRENDO PENINSULAR (Antilocapra americana peninsularis) POPULATION AND HABITAT VIABILITY ASSESSMENT FOR THE PENINSULAR PRONGHORN (Antilocapra americana peninsularis) La Paz, Baja California Sur, Mexico Centro de Investigaciones Biologicas del Noroeste, S.C. 16-18 de Noviembre de 1994 SECCION 6 EDUCACION AMBIENTAL ENVIRONMENTAL EDUCATION EDUCACION AMBIENTAL Campaiias de Difusion Hay que buscar un nombre para la campafia que la identifique, que Harne la atenci6n y que evoque ala subespecie, se propone: "Corre, berrendo, corre, la extinci6n es para siempre". Toda esta implementaci6n requerie de recursos financieros con instituciones interesadas: • SEP (Secretaria de Educaci6n Publica) • ESSA (Compafiia Exportadora de Sal, S.A.) • Prestadores de servicios turisticos • SEDESOL (Secretaria de Desarrollo Social) • CIBNOR (Centro de Investigaciones Biol6gicas del Noroeste, S.C.) • Grupos ecologicos • SARH (Secretaria de Agricultura y Recursos Hidniulicos) • Sierra Madre, S.C. • Unidos para la Conservaci6n, A.C. • CBSG (Conservation Breeding Specialist Group) • USFWS (United States Fish and Wildlife Service) • NAPF (North American Pronghorn Foundation) • PEMEX (Petroleos Mexicanos) • CFE (Comisi6n Federal de Electricidad) • WWF (World Wide Fund for Nature) • Comite Conjunto Mexico-Estados Unidos de America para la Conservaci6n de la Vida Silvestre • Conservaci6n Internacional • AGFD (Arizona Game and Fish Department) • CFGD (California Fish and Game Department) • Gobierno del estado de Baja California Sur • CONACYT (Consejo nacional de Ciencia y Tecnologia) • The Nature Conservancy • CONABIO (Comisi6n Nacional para el Conocimiento y Uso de la Biodiversidad) • National Science Foundation • National Wildlife Foundation • Arizona Bighorn Sheep Society • Wildlife Conservation Society • Wildlife Management Institute • National Geographic Society • SIMAC (Sistema de Investigaciones del mar de Cortes) Peninsular Pronghorn PHVA Report 63 Nivel Local, Regional, Nacional y Internacional Acciones en las poblaciones de la Reserva adyacentes ala distribuci6n del berrendo (Gro. Negro, Vizcaino, San Ignacio, Ejido, San Jose de Castro, etc.). Regional Principales centros poblacionales, cabeceras municipales y centros de interes turistico. Acciones a Corto Plazo • Poster: Puede ser patrocinado por un grupo intemacional, Sierra Madre, ESSA, SEDESOL, CFE, y Pemex. La mitad se distribuye para difusi6n, y la otra mitad se vende para obtener fondos que puede servir para sefializaci6n. • Sefializacion: En los principales caminos y berchas, de acuerdo al estudio de zonificaci6n que se derive del estudio de inspecci6n y vigilancia. • Campana en radio y peri6dico: Texto (ubicaci6n de la subespecie, importancia de la subespecie, (mica y en peligro y que es un emblema). • Capacitaci6n: A maestros, grupo de la REBIVI y vigilantes. • Audiovisual • Documental • Concursos (de dibujos y de otra indole) • Triptico ( elaborado por turismo) • Folleto didactico para diferentes niveles, que incluya uno para obtener fondos. • Publicaci6n de articulos en revistas de difusion (Escala etc.). Libro sobre la Reserva de la Biosfera • Spots • Timbre 64 Peninsular Pronghorn PHVA Report • Imagen en offset de 500 piezas que se venden a N$300.00. • Recursos para sefializaci6n por patrocinadores. Materiales con slogan para obtener fondos: camisetas, gorras, calendario de la Reserva. Peninsular Pronghorn PHVA Report 65 66 Peninsular Pronghorn PHVA Report ENVIRONMENTAL EDUCATION Advertising Campaign A slogan that identifies and brings attention to the subspecies: "Run, berrendo, run, extinction is forever". All of this implementation requires financial resources from all interested institutions: • SEP (Secretary of Public Education) • ESSA (Salt Mining and Export) • Tourism Agencies • SEDESOL (Secretary of Social Development) • CIBNOR (Center for Biological Investigation of the Northeast) • Ecological groups • SARH (Secretary of Agriculture and Hydrology) • Sierre Madre • United for Conservation • CBSG (Conservation Breeding Specialist Group) • USFWS (U.S. Fish & Wildlife Service) • NAPF (North American Pronghorn Foundation) • PEMEX (Mexican Petroleum) • CFE (Federal Electricity Commission) • WWF (World Wide Fund for Nature) • Joint Mexico-US Committee for Wildlife Conservation • Conservation International • AGFD (Arizona Game and Fish Department) • CFGD (California Fish and Game Department) • Government of the State of Baja California Sur • CONACYT (National Council for Science and Technology) • The Nature Conservancy • CONABIO (National Commission for the Understanding and Use of Biodiversity) • National Science Foundation • National Wildlife Foundation • Arizona Bighorn Sheep Society • Wildlife Conservation Society • Wildlife Management Institute • National Geographic Society • SIMAC (System for Research in the Sea of Cortes) Peninsular Pronghorn PHVA Report 67 Local, Regional, National, and International Levels Actions in those towns adjacent to the Reserve that are close to the pronghorn distribution: Guerrero Negro, Vizcaino, San Ignacio, Ejido, San Jose de Castro, etc. Regional Main population centers, municipalities, and tourist centers of interest. Short-term Actions • Poster: It can be sponsored by an international group such as Sierre Madre, ESSA, SEDESOL, CFE, or PEMEX. Half of the posters can be distributed, and the other half sold to obtain funds that can help further education efforts. • Sign production: On principal paved and unpaved roads, according to recent zoning studies which will result from studies on pronghorn distribution. • Newspaper and Radio Campaign: Text (location of the subspecies, importance of the subspecies, unique, endangered, local symbol). • Training: for scientists and managers of the Vizcaino Biosphere Reserve. • Audio-visual • Documentaries • Contests: for drawings and other items • Pamphlet: made by tourism agencies Educational brochures written on many levels, including one for obtaining funding. • Publication of articles in non-scientific journals (Escala, etc.). • Book about the Biosphere Reserve • Short television and/or radio public service announcements • Stamps 68 Peninsular Pronghorn PHVA Report • 500 offset images of the berrendo, sold at N$300.00 each • Resources for sign production by sponsors • Materials with slogans to obtain funding: T -shirts, caps, and calendars of the Reserve. Peninsular Pronghorn PHVA Report 69 70 Peninsular Pronghorn PHVA Report EVALUACION DE VIABILIDAD DE LA POBLACION Y DEL HABITAT DEL BERRENDO PENINSULAR (Antilocapra americana peninsularis) POPULATION AND HABITAT VIABILITY ASSESSMENT FOR THE PENINSULAR PRONGHORN (Antilocapra americana peninsularis) La Paz, Baja California Sur, Mexico Centro de Investigaciones Biologicas del Noroeste, S.C. 16-18 de Noviembre de 1994 SECCION 7 INVESTIGACION RESEARCH INVESTIGACION RESEARCH En la siguiente tabla (pagina 74) se enlistan las areas de estudio que es necesario abordar. Este listado no excluye ningun otro estudio (parcial o general) que no este contenido en el mismo. The following table (page 74) presents a list of studies needed. This list does not exclude any other study not explicitly considered. A 1 = corto plazo/short-term B 2 = mediano plazo/medium-term C 3 = largo plazo/long-term R = Nombres de los responsable/Names of those responsible: JCH Jorge Cancino Hernandez JDV James DeVos Jr. LC Laura Colton VSS Victor Sanchez Sotomayor L T Laura Thompson-Olay RRE Ricardo Rodriguez-Estrella RML Rodrigo Medellin Legorreta FRR Felipe Ramirez Ruiz de Velasco It is important to present detailed plans of every action that the group deems necessary. Funding agencies: V er lista en Educaci6n Ambiental See list in Environmental Education Peninsular Pronghorn PHVA Report 73 Prior. Calen. R Monitoreo sistematico y censos 1 A:3 JCH/JDV Systematic monitoring and surveying Biologia (demografia, requerimeientos de agua, dieta, 1-3 A:1-3 JCH/JDV parasitos, enfermedades) Life history (demography, water requirements, diet analysis, parasites, diseases) Taxonomia de la subespecie peninsular 3 B:2 LC/VSS/L T/JCH Taxonomy ofpeninsular subspecies Uso estacional del habitat, preferencias de habitat, 1 A:3 JCH/JDV identificaci6n de areas criticas Seasonal habitat use, habitat preferences, identifying critical areas (core areas) Evaluaci6n del habitat (hist6rico y presente, fragmentaci6n 1 A:3 RRE/JCH del habitat) Assessment ofthe suitable habitat (historical and present, habitat fragmentation) Eva1uaci6n y monitoreo de la calidad del habitat 1 A:2 RRE/JCH Evaluating and monitoring habitat quality Competencia, interacciones e impacto entre berrendo y 2 B:2 RMLIJCH ganado (habitos de alimentaci6n y disponibilidad de forreje, feno1ogia vegetal) Competition, interactions and impact between pronghorns and livestock (food habits and plant availability, plant phenology) Depredaci6n 1 A:2 RMLIJCH Predation Patrones climaticos 2 A:l RRE/LT Climatic patterns Evaluaci6n de la caceria furtiva 3 A:3 Evaluation ofpoaching Diversidad genetica y consanguinidad 1 A:l LC/JCH Genetic diversity and inbreeding Crianza en semicautiverio y reintroducci6n 1 B:3 VSS/FRR Fawn semi-captive rearing and reintroduction 74 Peninsular Pronghorn PHVA Report EVALUACION DE VIABILIDAD DE LA POBLACION Y DEL HABITAT DEL BERRENDO PENINSULAR (Antilocapra americana peninsularis) POPULATION AND HABITAT VIABILITY ASSESSMENT FOR THE PENINSULAR PRONGHORN (Antilocapra americana peninsularis) La Paz, Baja California Sur, Mexico Centro de Investigaciones Biologicas del Noroeste, S.C. 16-18 de Noviembre de 1994 SECCION 8 ESTRATEGIAS DE MANEJO MANAGEMENT STRATEGIES ESTRATEGIAS DE MANEJO Manejo en Cautiverio y Semicautiverio Tomando como base las experiencias del berrendo en cautiverio tanto en Mexico como en E.U., las cuales no tuvieron exito para alcanzar la sobrevivencia o adaptaci6n de las crias en estado silvestre se recomienda que el rnanejo de la mayor parte de la poblaci6n del berrendo peninsular en Vizcaino sea en estado silvestre. En este sentido se recornienda seguir las siguientes estrategias de manejo. Programa de Vigilancia Se requiere un programa, sectorizaci6n del area y en el corto plazo hacer enfasis en las areas de concentraci6n del berrendo la base sera de este prograrna el prograrna de vigilancia de la rnisma Reserva de la Biosfera "El Vizcaino". Normatividad del Uso del Suelo del Area de Distribucion Actual del Berrendo Tomando como base el ordenarniento de uso del suelo, se recornienda elaborar una norrnatividad especifica para el uso del suelo en el area del berrendo. Sefi.alizacion Se debe elaborar un prograrna de sefializaci6n especffico para el berrendo en congruencia con el prograrna de rnanejo de La Reserva. Tenencia de la Tierra y Uso del Suelo Se deben elaborar convenios con los duefios de los predios para lograr la protecci6n del berrendo, en compatibilidad con los prograrnas de conservacion de la subespecie, haciendo especial enfasis en el desarrollo de la ganaderia y agricultura. Gestionando con los diversos sectores que el desarrollo de obras de infraestructura se adecuen al ordenarniento de uso de suelo y vigilar su cumplimiento. Programa de Control de Depredadores Con base al resultados de las investigaciones en el area de podran llevar en el mediano y largo plazo. Peninsular Pronghorn PHVA Report 77 Manejo de Habitat Con base en los resultados de las investigaciones se podran llevar a cabo programas de restauraci6n de habitat. Estrategias de Manejo Prior. Calen. Resp. Inspecci6n y vigilancia (presencia en el area) 1 A:3 S,P Analisis y reglamentaci6n en el uso del suelo 1 B:2 S,C Gesti6n para coord. intersectoriales 1 A:3 c Control de depredadores 2* B:1-3 S,C Manejo de habitat 2* B:2-3 S,C Control y seguimiento de poblaci6n por 1 ** A:3 S,C,A telemetria Manejo en semicautiverio 1 *** B:3 B,C,E, s * Este punto queda sujeta a los resultados de las investigaciones ** 10 ejemplares para capturar. De acuerdo con la experiencia de los participantes de Arizona y Sonora, se recomienda inicialmente empezar con el marcado de 10 animales. *** La crianza en cautiverio entendida como la captura de crias y liberaci6n de subadultos esta sujeta ala preparaci6n de una propuesta que sera evaluada por un grupo de expertos en crianza en cautiverio y berrendo. La realizaci6n de este programa es independiente del comportamiento de la poblaci6n. 78 Peninsular Pronghorn PHVA Report MANAGEMENT STRATEGIES Captive and Semi-Captive Management Captive-rearing experiences with the pronghorn in Mexico and the U.S. have not resulted in survival or adaptation of fawns in the wild. Therefore, we recommend that management ofthe pronghorn in the Vizcaino Desert should be accomplished in the wild. The management strategies listed below should be followed: Vigilance Program The occupied habitat should be categorized. For the short term, law-enforcement efforts should be emphasized in areas of pronghorn concentration. The basis for this program is the vigilance program for the Vizcaino Biosphere Reserve. Land Use Regulations Based upon land use ordinances, it is recommended that specific land use legislation be developed in the pronghorn range. Radio Telemetry It is necessary to develop a specific radio telemetry program for the pronghorn in accordance with the management program of the Reserve. Land and Soil Use Legal agreements should be developed with landowners to achieve the protect the pronghorn and its habitat. Such agreements should encourage land management actions beneficial to the pronghorn. The agreements are particularly needed with landowners involved with agriculture or livestock grazing. Predator Control Program Based on results of research conducted in the area, predator control should be developed in the near future and may be required for the long term. Peninsular Pronghorn PHVA Report 79 Habitat Management Habitat restoration programs may be developed based on the results of research. Management Strategy Prior. Calen. Resp. Law Enforcement 1 A:3 S,P Analysis and land-use regulations 1 B:2 S,C Inter-institutional coordination 1 A:3 c Predator control 2* B:1-3 S,C Habitat management 2* B:2-3 S,C Population monitoring using telemetry 1** A:3 S,C,A Semi-captive management 1*** B:3 B,C,E,S * This strategy is dependent on results of research. ** 10 individuals to be captured. Based on experience from participants from Arizona and Sonora, it is recommended that the program be initiated with 10 animals. *** Rearing in captivity, defined as the capture of wild fawns and subsequent release of subadults, is subjected to the development of a proposal that will be evaluated by a group of experts in captive rearing and in pronghorn biology. Development of this program is independent of the demographic characteristics of the population. 80 Peninsular Pronghorn PHVA Report EVALUACION DE VIABILIDAD DE LA POBLACION Y DEL HABITAT DEL BERRENDO PENINSULAR (Antilocapra americana peninsularis) POPULATION AND HABITAT VIABILITY ASSESSMENT FOR THE PENINSULAR PRONGHORN (Antilocapra americana peninsularis) La Paz, Baja California Sur, Mexico Centro de Investigaciones Biol6gicas del Noroeste, S.C. 16-18 de Noviembre de 1994 SECCION 9 REFERENCIAS REFERENCES REFERENCIAS REFERENCES Barrett, M.W. 1984. Movements, habitat use, and predation on pronghorn fawns in Alberta. J. Wildl. Manage. 48:542-550. Cancino, J. 1988. Habitos de alimentaci6n del berrendo peninsular (Antilocapra americana peninsularis). Bachelor Thesis. Universidad Aut6noma Chapingo, Chapingo, Mexico. 66pp. Cancino, J. in press. Food habits of the peninsular pronghorn. Cancino, J., R. Rodriguez-Estrella, and A. Ortega. 1994. First aerial survey of historical range for peninsular pronghorn of Baja California, Mexico. J. Arizona-Nevada Acad. Sci 28:46-50. Diario Oficial. 1991. Acuerdo por el que se establecen los criterios ecol6gicos CT-CERN-001-91 que determinan las especies raras, amenazadas, en peligro de extinction o sujetas a protecci6n especial y sus endemismos, de la flora y fauna terrestres y acuaticas en la Republica Mexicana. Diario Oficial dele Federaci6n. Torno CDLII. No. 12. pp. 7-35. Hall, E.R. 1981. The Mammals ofNorth America. Vol. II. 2nd Edition. John Wiley and Sons, New York. Jaramillo, F. 1989. Contribuci6n al conomiciento y conservaci6n del berrendo de Baja California (Antilocapra americana peninsularis, Nelson 1912: Antilocapridae, Mammalia) en el Desierto de Vizcaino, Baja California Sur, Mexico. Tesis profesional. Universidad Nacional Aut6noma de Mexico. 140 pp. Nelson, E.W. 1925. Status ofthe pronghorned antelope, 1922-1924. Bull. No. 1346. U.S. Department of Agriculture, Washington, DC. 64pp. Ralls, K., J.D. Ballou, and A.R. Templeton. 1988. Estimates oflethal equivalents and the cost of inbreeding in mammals. Conservation Biology 2:185-193. Ricklefs, R.E. 1979. Ecology. 2nd Edition. Chiron, New York. SEDESOL. 1993. Secretaria de Desarrollo Social. Censo del berrendo de Baja California. Informe interno. Delegaci6n estatal en Baja California Sur. Peninsular Pronghorn PHVA Report 83 EVALUACION DE VIABILIDAD DE LA POBLACION Y DEL HABITAT DEL BERRENDO PENINSULAR (Antilocapra americana peninsularis) POPULATION AND HABITAT VIABILITY ASSESSMENT FORTHEPENINSULARPRONGHORN (Antilocapra americana peninsularis) La Paz, Baja California Sur, Mexico Centro de Investigaciones Biologicas del Noroeste, S.C. 16-18 de Noviembre de 1994 SECCION 10 LISTA DE PARTICIPANTES LIST OF PARTICIPANTS LISTA DE PARTICIPANTES- PARTICIPANTS Francisco Anguiano Sergio Alvarez Castaneda SEDESOL Centro de Investigaciones Biol6gicas del Calz. A. Olachea E/ Noroeste, S.C. Palacio de Justicia Apartado Postal128 La Paz, 23000, B.C.S. La Paz 23000, B.C.S. Mexico Tel: (112) 53633 Tel: (112) 21699 Fax: (112) 53625 Izquierdo Apolo Bioi. Carlos Castillo Pal. Gob. 2 Nivel Centro Ecol6gico de Sonora Gob. Edo. I. Carretera a Guaymas Km 2 La Catol. E/Allende y Apartado Postal 1497 23075 Bravo Guaymas, 83000, Sonora (112) 29477 Ext. 242 Tel: 62-50-1034-1137 Fax: 62-50-1236-1225 Benjamin Ballesteros SEDESOL Laura Colton Calz. A. Olachea E/ California Fish and Game Palacio de Justicia 1416 Ninth Street La Paz, 23000, B.C.S. Sacramento CA 95814 USA Mexico Tel: 916-653-6886 Tel: (112) 21699 James DeVos Jr. Carmen Blazquez Arizona Game and Fish Department Centro de Investigaciones Biol6gicas del 2222 West Greenway Road Noroeste, S.C. Phoenix AZ 85023 USA Apartado Postal128 Tel: (602) 942-3000 La Paz, 23000, B. C. S. Tel: (112) 53633 Patricia Gallina Fax: (112) 53625 Centro de Investigaciones Biol6gicas del Noroeste, S.C. Ing. Jorge Cancino Apartado Postal 128 Centro de Investigaciones Biol6gicas del La Paz 23000, B.C.S. Noroeste, S. C, Tel: (112) 53633 Apartado Postal128 Fax: (112) 53625 La Paz, 23000, B. C. S. Tel: (112) 53633 Fax: (112) 53625 Peninsular Pronghorn PHVA Report 87 Patricio Robles Gil Carlos Manterola Agrupaci6n Sierra Madre, S. C. Agrupaci6n Sierra Madre, S. C. Insurgentes Sur 949-701 Insurgentes Sur 949-701 Col. Napoles Col. Napoles Mexico, 03810, D. F. Mexico, 03810, D. F. Tel: (5) 682-8687 Tel: (5) 682-8687 Fax: (5) 520-8820 Fax: (5) 520-8820 Marco Antonio Giron Dr. Rodrigo Medellin SEDESOL Centro de Ecologia Calz. A. Olachea E/ U.N.A.M. Palacio de Justicia Apartado postal 70-275 La Paz, 23000, B.C.S. Ciudad Universitaria Mexico Mexico, 04510, D.F. Tel: (112) 21699 Tel: 5-550-5215 Fax: 5-548-5259 Fernando Heredia ESSA Dr. Phil Miller Av. Baja California SIN Conservation Breeding Specialist Group Gro. Negro, B.C.S. 12101 Johnny Cake Ridge Road Tel: (112) 70505 Ext. 351 Apple Valley, Minnesota 55124 USA Tel (612) 431-9325 Tim Hill Fax: (612) 432-2757 Minnesota Zoological Gardens 13000 Zoo Boulevard Fernando Jaramillo Monroy Apple Valley, Minnesota 55124 PG 7 Consultores, S.C. U.S.A. Apartado Postal 1000 Tel: (612) 431-9272 Tuxtla Gutierrez Fax: (612) 431-9300 29000 Chiapas, Mexico Tel/Fax: (961) 51097 Jim Lewis US Fish & Wildlife Service J. Cesar Peralta 500 Gold A venue SW ESSA Albuquerque, NM 87103 Apartado Postal 1 U.S.A. Gro. Negro 23940 B.C.S. Tel: 505-766-2914 Tel: (115) 70505 Fax: 505-766-8063 88 Peninsular Pronghorn PHVA Report Felipe Ramirez Jose A. Sanchez Zoologico Pachuca SEDESOL Tesoreros #30 Calz. A. 0 lachea E/ Col. Torriello Guerra Palacio de Justicia Tlalpan, 14050, Mexico D.F. C.P. La Paz, 23000, B.C.S. MEXICO Tel: (915) 8461039 Tel: (112) 21699 J. Carlos Ramirez Victor Sanchez Sotomavor SEDESOL Biocenosis, A.C. Calz. A. Olachea E/ Av. Universidad 1863 Palacio de Justicia Col Oxtopulco La Paz, 23000, B.C.S. Mexico, 04310, D.F. Mexico Tel: 5-661-9186 Tel: (112) 21699 Fax: 5-661-8699 Jose Maria Reyes Jose Bernal Stoopen Direcci6n General de Flora y Fauna Texas A&M University Silvestres. 240 East 47th St. #22C Av. Nuevo Leon 210 New York NY 10017 USA Mexico, D. F. Tel & Fax: (212) 688-9208 Tel: (5) 584-7578 Fax: (5) 574-8346 Gerardo Tapia-Hervert AZCARM M. en C. Ricardo Rodrigues-Estrella 11 Oriente 2407 Centro de Investigaciones Biol6gicas del Col. A viacion Noroeste, S.C. Puebla 72007, Pue. Apartado Postal128 Tel: (22) 358713/8 La Paz 23000, B.C.S. Fax: (22) 358607 Tel: (112) 53633 Fax: (112) 53625 Laura Thompson-Olais Cabeza Prieta National Wildlife Refuge Bernardo Sanabria 1611 North Second Avenue SEDESOL Ajo AZ 85321 USA Gemenis 14-B Col. Fovisste Tel: (602) 387-6483 Gro. Negro 23940 B.C.S. Fax: (602) 387-5359 Tel: (112) 71380 Gabriel Velazquez SEDESOL Calz. A. 0 lachea E/ Palacio de Justicia La Paz, 23000, B.C.S. Mexico Tel: (112) 21699 Peninsular Pronghorn PHVA Report 89 EVALUACION DE VIABILIDAD DE LA POBLACION Y DEL HABITAT DEL BERRENDO PENINSULAR (Antilocapra americana peninsularis) POPULATION AND HABITAT VIABILITY ASSESSl\1ENT FORTHEPENINSULARPRONGHORN (Antilocapra americana peninsularis) La Paz, Baja California Sur, Mexico Centro de Investigaciones Biologicas del Noroeste, S.C. 16-18 de Noviembre de 1994 SECCION 11 REFERENCIA TECNICA DE VORTEX VORTEX TECHNICAL REFERENCE VORTEX Technical Reference 91 Wildt. Res., 1993, 20, 45-65 VORTEX: A Computer Simulation Model for Population Viability Analysis Robert C. Lacy Department of Conservation Biology, Chicago Zoological Society, Brookfield, Illinois 60513, U.S.A. Abstract Population Viability Analysis (PVA) is the estimation of extinction probabilities by analyses that incorporate identifiable threats to population survival into models of the extinction process. Extrinsic forces, such as habitat loss, over-harvesting, and competition or predation by introduced species, often lead to population decline. Although the traditional methods of wildlife ecology can reveal such deterministic trends, random fluctuations that increase as populations become smaller can lead to extinction even of populations that have, on average, positive population growth when below carrying capacity. Computer simulation modelling provides a tool for exploring the viability of populations subjected to many complex, interacting deterministic and random processes. One such simulation model, VORTEX, has been used extensively by the Captive Breeding Specialist Group (Species Survival Commission, IUCN), by wildlife agencies, and by university classes. The algorithms, structure, assumptions and applications of VORTEX are described in this paper. VORTEX models population processes as discrete, sequential events, with probabilistic outcomes. VoRTEX simulates birth and death processes and the transmission of genes through the generations by generating random numbers to determine whether each animal lives or dies, to determine the number of progeny produced by each female each year, and to determine which of the two alleles at a genetic locus are transmitted from each parent to each offspring. Fecundity is assumed to be independent of age after an animal reaches reproductive age. Mortality rates are specified for each pre-reproductive age-sex class and for reproductive-age animals. Inbreeding depression is modelled as a decrease in viability in inbred animals. The user has the option of modelling density dependence in reproductive rates. As a simple model of density dependence in survival, a carrying capacity is imposed by a probabilistic truncation of each age class if the population size exceeds the specified carrying capacity. VoRTEX can model linear trends in the carrying capacity. VORTEX models environmental variation by sampling birth rates, death rates, and the carrying capacity from binomial or normal distributions. Catastrophes are modelled as sporadic random events that reduce survival and reproduction for one year. VoRTEX also allows the user to supplement or harvest the population, and multiple subpopulations can be tracked, with user-specified migration among the units. VoRTEX outputs summary statistics on population growth rates, the probability of population extinction, the time to extinction, and the mean size and genetic variation in extant populations. VoRTEX necessarily makes many assumptions. The model it incorporates is most applicable to species with low fecundity and long lifespans, such as mammals, birds and reptiles. It integrates the interacting effects of many of the deterministic and stochastic processes that have an impact on the viability of small populations, providing opportunity for more complete analysis than is possible by other techniques. PVA by simulation modelling is an important tool for identifying populations at risk of extinction, determining the urgency of action, and evaluating options for management. Introduction Many wildlife populations that were once widespread, numerous, and occupying con tiguous habitat, have been reduced to one or more small, isolated populations. The causes of the original decline are often obvious, deterministic forces, such as over-harvesting, 1035-3712/93/010045$10.00 92 VORTEX Technical Reference habitat destruction, and competition or predation from invasive introduced species. Even if the original causes of decline are removed, a small isolated population is vulnerable to additional forces, intrinsic to the dynamics of small populations, which may drive the population to extinction (Shaffer 1981; Soule 1987; Clark and Seebeck 1990). Of particular impact on small populations are stochastic processes. With the exception of aging, virtually all events in the life of an organism are stochastic. Mating, reproduction, gene transmission between generations, migration, disease and predation can be described by probability distributions, with individual occurrences being sampled from these distributions. Small samples display high variance around the mean, so the fates of small wildlife populations are often determined more by random chance than by the mean birth and death rates that reflect adaptations to their environment. Although many processes affecting small populations are intrinsically indeterminate, the average long-term fate of a population and the variance around the expectation can be studied with computer simulation models. The use of simulation modelling, often in con junction with other techniques, to explore the dynamics of small populations has been termed Population Viability Analysis (PV A). PVA has been increasingly used to help guide management of threatened species. The Resource Assessment Commission of Australia (1991) recently recommended that 'estimates of the size of viable populations and the risks of extinction under multiple-use forestry practices be an essential part of conservation planning'. Lindenmayer eta!. (1993) describe the use of computer modelling for PYA, and discuss the strengths and weaknesses of the approach as a tool for wildlife management. In this paper, I present the PVA program VORTEX and describe its structure, assumptions and capabilities. VORTEX is perhaps the most widely used PYA simulation program, and there are numerous examples of its application in Australia, the United States of America and elsewhere. The Dynamics of Small Populations The stochastic processes that have an impact on populations have been usefully categor ised into demographic stochasticity, environmental variation, catastrophic events and genetic drift (Shaffer 1981). Demographic stochasticity is the random fluctuation in the observed birth rate, death rate and sex ratio of a population even if the probabilities of birth and death remain constant. On the assumption that births and deaths and sex determination are stochastic sampling processes, the annual variations in numbers that are born, die, and are of each sex can be specified from statistical theory and would follow binomial distributions. Such demographic stochasticity will be important to population v:ability only in populations that are smaller than a few tens of animals (Goodman 1987), in which cases the annual frequencies of birth and death events and the sex ratios can deviate far from the means. The distribution of annual adult survival rates observed in the remnant population of whooping cranes (Grus americana) (Mirande eta!. 1993) is shown in Fig. 1. The innermost curve approximates the binomial distribution that describes the demographic stochasticity expected when the probability of survival is 92·7% (mean of 45 non-outlier years). Environmental variation is the fluctuation in the probabilities of birth and death that results from fluctuations in the environment. Weather, the prevalence of enzootic disease, the abundances of prey and predators, and the availability of nest sites or other required microhabitats can all vary, randomly or cyclically, over time. The second narrowest curve on Fig. I shows a normal distribution that statistically fits the observed frequency histogram of crane survival in non-outlier years. The difference between this curve and the narrower distribution describing demographic variation must be accounted for by environmental variation in the probability of adult survival. Catastrophic variation is the extreme of environmental variation, but for both method ological and conceptual reasons rare catastrophic events are analysed separately from the more typical annual or seasonal fluctuations. Catastrophes such as epidemic disease, VORTEX Technical Reference 93 0.15 >- c0 Q) ::> 0.10 0" ~ lL 0.05 0.00 0.5 0.6 0.7 0.8 0.9 1.0 Survival Fig. 1. Frequency histogram of the proportion of whooping cranes surviving each year, 1938-90. The broadest curve is the normal distribution that most closely fits the overall histogram. Statistically, this curve fits the data poorly. The second highest and second broadest curve is the normal distribution that most closely fits the histogram, excluding the five leftmost bars (7 outlier 'catastrophe' years). The narrowest and tallest curve is the normal approximation to the binomial distribution expected from demographic stochasticity. The difference between the tallest and second tallest curves is the variation in annual survival due to environmental variation. hurricanes, large-scale fires, and floods are outliers in the distribution of environmental variation (e.g. five leftmost bars on Fig. 1). As a result, they have quantitatively and sometimes qualitatively different impacts on wildlife populations. (A forest fire is not just a very hot day.) Such events often precipitate the final decline to extinction (Simberloff 1986, 1988). For example, one of two populations of whooping crane was decimated by a hurricane in 1940 and soon after went extinct (Doughty 1989). The only remaining population of the black-footed ferret (Mustela nigripes) was being eliminated by an outbreak of distemper when the last 18 ferrets were captured (Clark 1989). Genetic drift is the cumulative and non-adaptive fluctuation in allele frequencies resulting from the random sampling of genes in each generation. This can impede the recovery or accelerate the decline of wildlife populations for several reasons (Lacy 1993). Inbreeding, not strictly a component of genetic drift but correlated with it in small populations, has been documented to cause loss of fitness in a wide variety of species, including virtually all sexually reproducing animals in which the effects of inbreeding have been carefully studied (Wright 1977; Falconer 1981; O'Brien and Evermann 1988; Ralls et a!. 1988; Lacy et a!. 1993). Even if the immediate loss of fitness of inbred individuals is not large, the loss of genetic variation that results from genetic drift may reduce the ability of a population to adapt to future changes in the environment (Fisher 1958; Robertson 1960; Selander 1983). Thus, the effects of genetic drift and consequent loss of genetic variation in individuals and populations have a negative impact on demographic rates and increase susceptibility to environmental perturbations and catastrophes. Reduced population growth and greater fluctuations in numbers in turn accelerate genetic drift (Crow and Kimura 1970). These synergistic destabilising effects of stochastic process on small populations of wildlife have been described as an 'extinction vortex' (Gilpin and Soule 1986). The size below which a population is likely to be drawn into an extinction vortex can be considered a 'minimum 94 VORTEX Technical Reference viable population' (MVP) (Seal and Lacy 1989), although Shaffer (1981) first defined a MVP more stringently as a population that has a 99% probability of persistence for 1000 years. The estimation of MVPs or, more generally, the investigation of the probability of extinction constitutes PYA (Gilpin and Soule 1986; Gilpin 1989; Shaffer 1990). Methods for Analysing Population Viability An understanding of the multiple, interacting forces that contribute to extinction vortices is a prerequisite for the study of extinction-recolonisation dynamics in natural populations inhabiting patchy environments (Gilpin 1987), the management of small populations (Clark and Seebeck 1990), and the conservation of threatened wildlife (Shaffer 1981, 1990; Soule 1987; Mace and Lande 1991). Because demographic and genetic processes in small populations are inherently unpredictable, the expected fates of wildlife populations can be described in terms of probability distributions of population size, time to extinction, and genetic variation. These distributions can be obtained in any of three ways: from analytical models, from empirical observation of the fates of populations of varying size, or from simulation models. As the processes determining the dynamics of populations are multiple and complex, there are few analytical formulae for describing the probability distributions (e.g. Goodman 1987; Lande 1988; Burgmann and Gerard 1990). These models have incorporated only few of the threatening processes. No analytical model exists, for example, to describe the combined effect of demographic stochasticity and loss of genetic variation on the probability of population persistence. A few studies of wildlife populations have provided empirical data on the relationship between population size and probability of extinction (e.g. Belovsky 1987; Berger 1990; Thomas 1990), but presently only order-of-magnitude estimates can be provided for MVPs of vertebrates (Shaffer 1987). Threatened species are, by their rarity, unavailable and inappropriate for the experimental manipulation of population sizes and long-term monitoring of undisturbed fates that would be necessary for precise empirical measurement of MVPs. Retrospective analyses will be possible in some cases, but the function relating extinction probability to population size will differ among species, localities and times (Lindenmayer et at. 1993). Modelling the Dynamics of Small Populations Because of the lack of adequate empirical data or theoretical and analytical models to. allow ·prediction of the dynamics of populations of threatened species, various biologists have turned to Monte Carlo computer simulation techniques for PV A. By randomly sampling from defined probability distributions, computer programs can simulate the multiple, interacting events that occur during the lives of organisms and that cumulatively determine the fates of populations. The focus is on detailed and explicit modelling of the forces impinging on a given population, place, and time of interest, rather than on delineation of rules (which may not exist) that apply generally to most wildlife populations. Computer programs available to PYA include SPGPC (Grier 1980a, 1980b), GAPPS (Harris eta/. 1986), RAMAS (Ferson and Ak<;akaya 1989; Ak<;akaya and Ferson 1990; Ferson 1990), FORPOP (Possingham eta/. 1991), ALEX (Possingham eta/. 1992), and SIMPOP (Lacy eta/. 1989; Lacy and Clark 1990) and its descendant VORTEX. SIMPOP was developed in 1989 by converting the algorithms of the program SPGPC (written by James W. Grier of North Dakota State University) from BASIC to the c programming language. SIMPOP was used first in a PVA workshop organised by the Species Survival Commission's Captive Breeding Specialist Group (IUCN), the United States Fish and Wildlife Service, and the Puerto Rico Department of Natural Resources to assist in planning and assessing recovery efforts for the Puerto Rican crested toad (Peltophryne lemur). SIMPOP was subsequently used in PV A modelling of other species threatened VORTEX Technical Reference 95 with extinction, undergoing modification with each application to allow incorporation of additional threatening processes. The simulation program was renamed VORTEX (in reference to the extinction vortex) when the capability of modelling genetic processes was implemented in 1989. In 1990, a version allowing modelling of multiple populations was briefly named voRTICES. The only version still supported, with all capabilities of each previous version, is VORTEX Version 5.1. VORTEX has been used in PYA to help guide conservation and management of many species, including the Puerto Rican parrot (Amazona vittata) (Lacy et at. 1989), the Javan rhinoceros (Rhinoceros sondaicus) (Seal and Foose 1989), the Florida panther (Felis concolor coryi) (Seal and Lacy 1989), the eastern barred bandicoot (Perameles gunnii) (Lacy and Clark 1990; Maguire et at. 1990), the lion tamarins (Leontopithecus rosaiia ssp.) (Seal et at. 1990), the brush-tailed rock-wallaby (Petrogale penicillata penicillata) (Hill 1991), the mountain pygmy-possum (Burramys parvus), Leadbeater's possum (Gymnobelideus leadbeateri), the long-footed potoroo (Potorous longipes), the orange-bellied parrot (Neophema chrysogaster) and the helmeted honeyeater (Lichenostomus melanops cassidix) (Clark et at. 1991), the whooping crane (Crus americana) (Mirande et at. 1993), the Tana River crested mangabey (Cercocebus galeritus galeritus) and the Tana River red colobus (Colobus badius rufomitratus) (Seal eta!. 1991), and the black rhinoceros (Diceros bicornis) (Foose et at. 1992). In some of these PV As, modelling with VORTEX has made clear the insufficiency of past management plans to secure the future of the sp~cies, and alternative strategies were proposed, assessed and implemented. For example, the multiple threats to the Florida panther in its existing habitat were recognised as probably insurmountable, and a captive breeding effort has been initiated for the purpose of securing the gene pool and providing animals for release in areas of former habitat. PVA modelling with VORTEX has often identified a single threat to which a species is particularly vulnerable. The small but growing population of Puerto Rican parrots was assessed to be secure, except for the risk of population decimation by hurricane. Recommendations were made to make available secure shelter for captive parrots and to move some of the birds to a site distant from the wild flock, in order to minimise the damage that could occur in a catastrophic storm. These recommended actions were only partly implemented when, in late 1989, a hurricane killed many of the wild parrots. The remaining population of about 350 Tana River red co lobus were determined by PVA to be so fragmented that demographic and genetic processes within the lO subpopulations destabilised population dynamics. Creation of habitat corridors may be necessary to prevent extinction of the taxon. In some cases, PV A modelling has been reassuring to managers: analysis of black rhinos in Kenya indicated that many of the populations within sanctuaries were recovering steadily. Some could soon be used to provide animals for re-establishment or supplementation of populations previously eliminated by poaching. For some species, available data were insufficient to allow definitive PV A with VORTEX. In such cases, the attempt at PVA modelling has made apparent the need for more data on population trends and processes, thereby helping to justify and guide research efforts. Description of VORTEX Overview The VORTEX computer simulation model is a Monte Carlo simulation of the effects of deterministic forces, as well as demographic, environmental and genetic stochastic events, on wildlife populations. VORTEX models population dynamics as discrete, sequential events that occur according to probabilities that are random variables, following user-specified distributions. The input parameters used by VORTEX are summarised in the first part of the sample output given in the Appendix. VORTEX simulates a population by stepping through a series of events that describe an annual cycle of a typical sexually reproducing, diploid organism: mate selection, 96 VORTEX Technical Reference reproduction, mortality, increment of age by one year, migration among populations, removals, supplementation, and then truncation (if necessary) to the carrying capacity. The program was designed to model long-lived species with low fecundity, such as mammals, birds and reptiles. Although it could and has been used in modelling highly fecund vertebrates and invertebrates, it is awkward to use in such cases as it requires complete specification of the percentage of females producing each possible clutch size. Moreover, computer memory limitations often hamper such analyses. Although VORTEX iterates life events on an annual cycle, a user could model 'years' that are other than 12 months' duration. The simulation of the population is itself iterated to reveal the distribution of fates that the population might experience. Demographic Stochasticity VORTEX models demographic stochasticity by determining the occurrence of probabilistic events such as reproduction, litter size, sex determination and death with a pseudo-random number generator. The probabilities of mortality and reproduction are sex-specific and pre-determined for each age class up to the age of breeding. It is assumed that reproduction and survival probabilities remain constant from the age of first breeding until a specified upper limit to age is reached. Sex ratio at birth is modelled with a user-specified constant probability of an offspring being male. For each life event, if the random value sampled from the uniform 0-1 distribution falls below the probability for that year, the event is deemed to have occurred, thereby simulating a binomial process. The source code used to generate random numbers uniformly distributed between 0 and 1 was obtained from Maier (1991), according to the algorithm of Kirkpatrick and Stoll (1981). Random deviates from binomial distributions, with mean p and standard deviation s, are obtained by first determining the integral number of binomial trials, N, that would produce the value of s closest to the specified value, according to 2 N=p(l-p)!s • N binomial trials are then simulated by sampling from the uniform 0-l distribution to obtain the desired result, the frequency or proportion of successes. If the value of N determined for a desired binomial distribution is larger than 25, a normal approximation is used in place of the binomial distribution. This normal approximation must be truncated at 0 and at 1 to allow use in defining probabilities, although, with such large values of N, s is small relative to p and the truncation would be invoked only rarely. To avoid introducing bias with this truncation, the normal approximation to the binomial (when used) is truncated symmetrically around the mean. The algorithm for generating random numbers from a unit normal distribution follows Latour (1986). VORTEX can model monogamous or polygamous mating systems. In a monogamous system, a relative scarcity of breeding males may limit reproduction by females. In poly gamous or monogamous models, the user can specify the proportion of the adult males in the breeding pool. Males are randomly reassigned to the breeding pool each year of the simulation, and all males in the breeding pool have an equal chance of siring offspring. The 'carrying capacity', or the upper limit for population size within a habitat, must be specified by the user. VORTEX imposes the carrying capacity via a probabilistic truncation whenever the population exceeds the carrying capacity. Each animal in the population has an equal probability of being removed by this truncation. Environmental Variation VORTEX can model annual fluctuations in birth and death rates and in carrying capacity as might result from environmental variation. To model environmental variation, each VORTEX Technical Reference 97 demographic parameter is assigned a distribution with a mean and standard deviation that is specified by the user. Annual fluctuations in probabilities of reproduction and mortality are modelled as binomial distributions. Environmental variation in carrying capacity is modelled as a normal distribution. The variance across years in the frequencies of births and deaths resulting from the simulation model (and in real populations) will have two components: the demographic variation resulting from a binomial sampling around the mean for each year, and additional fluctuations due to environmental variation and catastrophes (see Fig. 1 and section on The Dynamics of Small Populations, above). Data on annual variations in birth and death rates are important in determining the probability of extinction, as they influence population stability (Goodman 1987). Unfor tunately, such field information is rarely available (but see Fig. 1). Sensitivity testing, the examination of a range of values when the precise value of a parameter is unknown, can help to identify whether the unknown parameter is important in the dynamics of a population. Catastrophes Catastrophes are modelled in VORTEX as random events that occur with specified probabilities. Any number of types of catastrophes can be modelled. A catastrophe will occur if a randomly generated number between zero and one is less than the probability of occurrence. Following a catastrophic event, the chances of survival and successful breeding for that simulated year are multiplied by severity factors. For example, forest fires might occur once in 50 years, on average, killing 25% of animals, and reducing breeding by survivors by 50% for the year. Such a catastrophe would be modelled as a random event with 0·02 probability of occurrence each year, and severity factors of 0·75 for survival and 0· 50 for reproduction. Genetic Processes Genetic drift is modelled in VORTEX by simulation of the transmission of alleles at a hypothetical locus. At the beginning of the simulation, each animal is assigned two unique alleles. Each offspring is randomly assigned one of the alleles from each parent. Inbreeding depression is modelled as a loss of viability during the first year of inbred animals. The impacts of inbreeding are determined by using one of two models available within VORTEX: a Recessive Lethals model or a Heterosis model. In the Recessive Lethals model, each founder starts with one unique recessive lethal allele and a unique, dominant non-lethal allele. This model approximates the effect of inbreeding if each individual in the starting population had one recessive lethal allele in its genome. The fact that the simulation program assumes that all the lethal alleles are at the same locus has a very minor impact on the probability that an individual will die because of homozygosity for one of the lethal alleles. In the model, homozygosity for different lethal alleles are mutually exclusive events, whereas in a multilocus model an individual could be homozygous for several lethal alleles simultaneously. By v,irtue of the death of individuals that are homozygous for lethal alleles, such alleles would be removed slowly by natural selection during the generations of a simulation. This reduces the genetic variation present in the population relative to the case with no inbreeding depression, but also diminishes the subsequent probability that inbred individuals will be homozygous for a lethal allele. This model gives an optimistic reflection of the impacts of inbreeding on many species, as the median number of lethal equivalents per diploid genome observed for mammalian populations is about three (Ralls et a!. 1988). The expression of fully recessive deleterious alleles in inbred organisms is not the only genetic mechanism that has been proposed as a cause of inbreeding depression. Some or 98 VORTEX Technical Reference most of the effects of inbreeding may be a consequence of superior fitness of heterozygotes (heterozygote advantage or 'heterosis'). In the Heterosis model, all homozygotes have reduced fitness compared with heterozygotes. Juvenile survival is modelled according to the logarithmic model developed by Morton et at. (1956): lnS=A -BF in which S is survival, F is the inbreeding coefficient, A is the logarithm of survival in the absence of inbreeding, and B is a measure of the rate at which survival decreases with inbreeding. B is termed the number of 'lethal equivalents' per haploid genome. The number of lethal equivalents per diploid genome, 2B, estimates the number of letha! alleles per individual in the population if all deleterious effects of inbreeding were due to recessive lethal alleles. A population in which inbreeding depression is one lethal equivalent per diploid genome may have one recessive lethal allele per individual (as in the Recessive Lethals model, above), it may have two recessive alleles per individual, each of which confer a 50% decrease in survival, or it may have some other combination of recessive deleterious alleles that equate in effect with one lethal allele per individual. Unlike the situation with fully recessive deleterious alleles, natural selection does not remove deleterious alleles at heterotic loci because all alleles are deleterious when homozygous and beneficial when present in heterozygous combination with other alleles. Thus, under the Heterosis model, the impact of inbreeding on survival does not diminish during repeated generations of inbreeding. Unfortunately, for relatively few species are data available to allow estimation of the effects of inbreeding, and the magnitude of these effects varies considerably among species (Falconer 1981; Ralls eta!. 1988; Lacy eta!. 1993). Moreover, whether a Recessive Lethals model or a Heterosis model better describes the underlying mechanism of inbreeding depression and therefore the response to repeated generations of inbreeding is not well known (Brewer eta!. 1990), and could be determined empirically only from breeding studies that span many generations. Even without detailed pedigree data from which to estimate the number of lethal equivalents in a population and the underlying nature of the genetic load (recessive alleles or heterosis), applications of PVA must make assumptions about the effects of inbreeding on the population being studied. In some cases, it might be considered appropriate to assume that an inadequately studied species would respond to inbreeding in' accord with the median (3 ·14 lethal equivalents per diploid) reported in the survey by Ralls et a!. (1988). In other cases, there might be reason to make more optimistic assumptions (perhaps the lower quartile, 0 · 90 lethal equivalents), or more pessimistic assumptions (perhaps the upper quartile, 5 · 62 lethal equivalents). Deterministic Processes VORTEX can incorporate several deterministic processes. Reproduction can be specified to be density-dependent. The function relating the proportion of adult females breeding each year to the total population size is modelled as a. fourth-order polynomial, which can provide a close fit to most plausible density-dependence curves. Thus, either positive population responses to low-density or negative responses (e.g. Allee effects), or more complex relationships, can be modelled. Populations can be supplemented or harvested for any number of years in each simulation. Harvest may be culling or removal of animals for translocation to another (unmodelled) population. The numbers of additions and removals are specified according to the age and sex of animals. Trends in the carrying capacity can also be modelled in VORTEX, specified as an annual percentage change. These changes are modelled as linear, rather than geometric, increases or decreases. VORTEX Technical Reference 99 Migration among Populations VORTEX can model up to 20 populations, with possibly distinct population parameters. Each pairwise migration rate is specified as the probability of an individual moving from one population to another. This probability is independent of the age and sex. Because of between-population migration and managed supplementation, populations can be recolonised. VORTEX tracks the dynamics of local extinctions and recolonisations through the simulation. Output VORTEX outputs (1) probability of extinction at specified intervals (e.g., every 10 years during a 100-year simulation), (2) median time to extinction if the population went extinct in at least 50% of the simulations, (3) mean time to extinction of those simulated popu lations that became extinct, and (4) mean size of, and genetic variation within, extant populations (see Appendix and Lindenmayer et at. 1993). Standard deviations across simulations and standard errors of the mean are reported for population size and the measures of genetic variation. Under the assumption that extinction of independently replicated populations is a binomial process, the standard error of the probability of extinction (SE) is reported by VORTEX as SE(p)=...f[px(l-p)ln], in which the frequency of extinction was p over n simulated populations. Demographic and genetic statistics are calculated and reported for each subpopulation and for the metapopulation. Availability of the VORTEX Simulation Program VORTEX Version 5.1 is written in the C programming language and compiled with the Lattice 80286C Development System (Lattice Inc.) for use on microcomputers using the MS-DOS (Microsoft Corp.) operating system. Copies of the compiled program and a manual for its use are available for nominal distribution costs from the Captive Breeding Specialist Group (Species Survival Commission, IUCN), 12101 Johnny Cake Ridge Road, Apple Valley, Minnesota 55124, U.S.A. The program has been tested by many workers, but cannot be guaranteed to be error-free. Each user retains responsibility for ensuring that the program does what is intended for each analysis. Sequence of Program Flow (1) The seed for the random number generator is initialised with the number of seconds elapsed since the beginning of the 20th century. (2) The user is prompted for input and output devices, population parameters, duration of simulation, and number of interations. (3) The maximum allowable population size (necessary for preventing memory over flow) is calculated as Nmax= (K + 3s) X (1 + L) in which K is the maximum carrying capacity (carrying capacity can be specified to change linearly for a number of years in a simulation, so the maximum carrying capacity can be greater than the initial carrying capacity), s is the annual environmental variation in the carrying capacity expressed as a standard deviation, and L is the specified maximum litter size. It is theoretically possible, but very unlikely, that a simulated population will exceed the calculated Nmax· If this occurs then the program will give an error message and abort. 100 VORTEX Technical Reference (4) Memory is allocated for data arrays. If insufficient memory is available for data arrays then Nmax is adjusted downward to the size that can be accommodated within the available memory and a warning message is given. In this case it is possible that the analysis may have to be terminated because the simulated population exceeds Nmax- Because Nmax is often several-fold greater than the likely maximum population size in a simulation, a warning it has been adjusted downward because of limiting memory often will not hamper the analyses. Except for limitations imposed by the size of the computer memory (VORTEX can use extended memory, if available), the only limit to the size of the analysis is that no more than 20 populations exchanging migrants can be simulated. (5) The expected mean growth rate of the population is calculated from mean birth .and death rates that have been entered. Algorithms foliow cohort life-table analyses (Ricklefs 1979). Generation time and the expected stable age distribution are also estimated. Life table estimations assume no limitation by carrying capacity, no limitation of mates, and no loss of fitness due to inbreeding depression, and the estimated intrinsic growth rate assumes that the population is at the stable age distribution. The effects of catastrophes are incorporated into the life-table analysis by using birth and death rates that are weighted averages of the values in years with and without catastrophes, weighted by the probability of a catastrophe occurring or not occurring. (6) Iterative simulation of the population proceeds via steps 7-26 below. For exploratory modelling, 100 iterations are usually sufficient to reveal gross trends among sets of simu lations with different input parameters. For more precise examination of population behaviour under various scenarios, 1000 or more simulations should be used to minimise standard errors around mean results. (7) The starting population is assigned an age and sex structure. The user can specify the exact age-sex structure of the starting population, or can specify an initial population size and request that the population be distributed according to the stable age distribution calculated from the life table. Individuals in the starting population are assumed to be unrelated. Thus, inbreeding can occur only in second and later generations. (8) Two unique alleles at a hypothetical genetic locus are assigned to each individual in the starting population and to each individual supplemented to the population during the simulation. VoRTEX therefore uses an infinite alleles model of genetic variation. The subsequent fate of genetic variation is tracked by reporting the number of extant alleles each year, the expected heterozygosity or gene diversity, and the observed heterozygosity. The expected heterozygosity, derived from the Hardy-Weinberg equilibrium, is given by in which Pi is the frequency of allele i in the population. The observed heterozygosity is simply the proportion of the individuals in the simulated population that are heterozygous. Because of the starting assumption of two unique alleles per founder, the initial population has an observed heterozygosity of 1 · 0 at the hypothetical locus and only inbred animals can become homozygous. Proportional loss of heterozygosity by means of random genetic drift is independent of the initial heterozygosity and allele fre9uencies of a population (assuming that the initial value was not zero) (Crow and Kimura 1970), so the expected heterozygosity remaining in a simulated population is a useful metric of genetic decay for comparison across scenarios and populations. The mean observed heterozygosity reported by VORTEX is the mean inbreeding coefficient of the population. (9) The user specifies one of three options for modelling the effect of inbreeding: (a) no effect of inbreeding on fitness, that is, all alleles are selectively neutral, (b) each founder individual has one unique lethal and one unique non-lethal allele (Recessive Lethals option), or (c) first-year survival of each individual is exponentially related to its inbreeding coefficient (Heterosis option). The first case is clearly an optimistic one, as almost all diploid VORTEX Technical Reference 101 populations studied intensively have shown deleterious effects of inbreeding on a variety of fitness components (Wright 1977; Falconer 1981). Each of the two models of inbreeding depression may also be optimistic, in that inbreeding is assumed to have an impact only on first-year survival. The Heterosis option allows, however, for the user to specify the severity of inbreeding depression on juvenile survival. (10) Years are iterated via steps 11-25 below. (ll) The probabilities of females producing each possible litter size are adjusted to account for density dependence of reproduction (if any). (12) Birth rate, survival rates and carrying capacity for the year are adjusted to model environmental variation. Environmental variation is assumed to follow binomial distributions for birth and death rates and a normal distribution for carrying capacity, with mean rates and standard deviations specified by the user. At the outset of each year a random number is drawn from the specified binomial distribution to determine the percentage of females producing litters. The distribution of litter sizes among those females that do breed is main tained constant. Another random number is drawn from a specified Linomial distribution to model the environmental variation in mortality rates. If environmental variations in reproduction and mortality are chosen to be correlated, the random number used to specify mortality rates for the year is chosen to be the same percentile of its binomial distribution as was the number used to specify reproductive rate. Otherwise, a new random number is drawn to specify the deviation of age- and sex-specific mortality rates for their means. Environmental variation across years in mortality rates is always forced to be correlated among age and sex classes. The carrying capacity (K) of the year is determined by first increasing or decreasing the carrying capacity at year I by an amount specified by the user to account for linear changes over time. Environmental variation in K is then imposed by drawing a random numbeF from a normal distribution with the specified values for mean and standard deviation. (13) Birth rates and survival rates for the year are adjusted to model any catastrophes determined to have occurred in that year. (14) Breeding males are selected for the year. A male of breeding age is placed into the pool of potential breeders for that year if a random number drawn for that male is less than the proportion of breeding-age males specified to be breeding. (15) For each female of breeding age, a mate is drawn at random from the pool of breeding males for that year. The size of the litter produced by that pair is determined by comparing the probabilities of each potential litter size (including litter size of 0, no breeding) to a randomly drawn number. The offspring are produced and assigned a sex by comparison of a random number to the specified sex ratio at birth. Offspring are assigned, at random, one allele at the hypothetical genetic locus from each parent. (16) If the Heterosis option is chosen for modelling inbreeding depression, the genetic kinship of each new offspring to each other living animal in the population is determined. The kinship between a new animal, A, and another existing animal, B is fAs=0·5 X(jMs+ fPB) in which fu is the kinship between animals i and j, M is the mother of A, and P is the father of A. The inbreeding coefficient of each animal is equal to the kinship between its parents, F=iMP• and the kinship of an animal to itself is fAA= 0· 5 X (1 +F). [See Ballou (1983) for a detailed description of this method for calculating inbreeding coefficients.] (17) The survival of each animal is determined by comparing a random number to the survival probability for that animal. In the absence of inbreeding depression, the survival probability is given by the age and sex-specific survival rate for that year. If the Heterosis model of inbreeding depression is used and an individual is inbred, the survival probability is multiplied by e-bF in which b is the number of lethal equivalents per haploid genome. 102 VORTEX Technical Reference If the Recessive Lethals model is used, all offspring that are homozygous for a lethal allele are killed. (18) The age of each animal is incremented by 1, and any animal exceeding the maximum age is killed. (19) If more than one population is being modelled, migration among populations occurs stochastically with specified probabilities. (20) If population harvest is to occur that year, the number of harvested individuals of each age and sex class are chosen at random from those available and removed. If the number to be removed do not exist for an age-sex class, voRTEX continues but reports that harvest was incomplete. (21) Dead animals are removed from the computer memory to make space for future generations. (22) If population supplementation is to occur in a particular year, new individuals of the specified age class are created. Each immigrant is assigned two unique alleles, one of which will be a recessive lethal in the Recessive Lethals model of inbreeding depression. Each immigrant is assumed to be genetically unrelated to all other individuals in the population. (23) The population growth rate is calculated as the ratio of the population size in the current year to the previous year. (24) If the population size (N) exceeds the carrying capacity (K) for that year, additional mortality is imposed across all age <~nd sex classes. The probability of each animal dying during this carrying capacity truncation is set to (N-K)lN, so that the expected population size after the additional mortality is K. (25) Summary statistics on population size and genetic variation are tallied and reported. A simulated population is determined to be extinct if one of the sexes has no representatives. (26) Final population size and genetic variation are determined for the simulation. (27) Summary statistics on population size, genetic variation, probability of extinction, and mean population growth rate, are calculated across iterations and printed out. Assumptions Underpinning VORTEX It is impossible to simulate the complete range of complex processes that can have an impact on wild populations. As a result there are necessarily a range of mathematical and biological assumptions that underpin any PVA program. Some of the more important assumptions in VORTEX include the following. (1) Survival probabilities are density independent when population size is less than carrying capacity. Additional mortality imposed when the population exceeds K affects all age and sex classes equally. (2) The relationship between changes in population size and genetic variability are examined for only one locus. Thus, potentially complex interactions between genes located on the same chromosome (linkage disequilibrium) are ignored. Such interactions are typically associated with genetic drift in very small populations, but it is unknown if, or how, they would affect population viability. (3) All animals of reproductive age have an equal probability of breeding. This ignores the likelihood that some animals within a population may have a greater probability of breeding successfully, and breeding more often, than other individuals. If breeding is not at random among those in the breeding pool, then decay of genetic variation and inbreeding will occur more rapidly than in the model. VORTEX Technical Reference 103 (4) The life-history attributes of a population (birth, death, migration, harvesting, supplementation) are modelled as a sequence of discrete and therefore seasonal events. How ever, such events are often continuous through time and the model ignores the possibility that they may be aseasonal or only partly seasonal. (5) The genetic effects of inbreeding on a population are determined in VORTEX by using one of two possible models: the Recessive Lethals model and the Heterosis model. Both models have attributes likely to be typical of some populations, but these may vary within and between species (Brewer et at. 1990). Given this, it is probable that the impacts of inbreeding will fall between the effects of these two models. Inbreeding is assumed to depress only one component of fitness: first-year survival. Effects on reproduction could be incorporated into this component, but longer-term impacts such as increased disease susceptibility or decreased ability to adapt to environmental change are not modelled. (6) The probabilities of reproduction and mortality are constant from the age of first breeding until an animal reaches the maximum longevity. This assumes that animals continue to breed until they die. (7) A simulated catastrophe will have an effect on a population only in the year that the event occurs. (8) Migration rates among populations are independent of age and sex. (9) Complex, interspecies interactions are not modelled, except in that such community dynamics might contribute to random environmental variation in demographic parameters. For example, cyclical fluctuations caused by predator-prey interactions cannot be modelled by VORTEX. Discussion Uses and Abuses of Simulation Modelling for PVA Computer simulation modelling is a tool that can allow crude estimation of the prob ability of population extinction, and the mean population size and amount of genetic diversity, from data on diverse interacting processes. These processes are too complex to be integrated intuitively and no analytic solutions presently, or are likely to soon, exist. PV A modelling focuses on the specifics of a population, considering the particular habitat, threats, trends, and time frame of interest, and can only be as good as the data and the assumptions input to the model (Lindenmayer et a!. 1993). Some aspects of population dynamics are not modelled by VORTEX nor by any other program now available. In particular, models of single-species dynamics, such as VORTEX, are inappropriate for use on species whose fates are strongly determined by interactions with other species that are in turn undergoing complex (and perhaps synergistic) population dynamics. Moreover, VORTEX does not model many conceivable and perhaps important interactions among variables. For example, loss of habitat might cause secondary changes in reproduction, mortality, and migration rates, but ongoing trends in these parameters cannot be simulated with VORTEX. It is important to stress that PVA does not predict in general what will happen to a population; PVA forecasts the likely effects only of those factors incorporated into the model. Yet, the use of even simplified computer models for PV A can provide more accurate predictions about population dynamics than the even more crude techniques available previously, such as calculation of expected population growth rates from life tables. For the purpose of estimating extinction probabilities, methods that assess only deterministic factors are almost certain to be inappropriate, because populations near extinction will commonly be so small that random processes dominate deterministic ones. The suggestion by Mace and Lande (1991) that population viability be assessed by the application of simple rules (e.g., a taxon be considered Endangered if the total effective population size is below 50 or the 1 04 VORTEX Technical Reference total census size below 250) should be followed only if knowledge is insufficient to allow more accurate quantitative analysis. Moreover, such preliminary judgments, while often important in stimulating appropriate corrective measures, should signal, not obviate, the need for more extensive investigation and analysis of population processes, trends and threats. Several good population simulation models are available for PVA. They differ in capabilities, assumptions and ease of application. The ease of application is related to the number of simplifying assumptions and inversely related to the flexibility and power of the model. It is unlikely that a single or even a few simulation models will be appropriate for all PV As. The VORTEX program has some capabilities not found in many other population simulation programs, but is not as flexible as are some others (e.g., GAPPS; Harris eta!. 1986). VORTEX is user-friendly and can be used by those with relatively little understanding of population biology and extinction processes, which is both an advantage and a disadvantage. Testing Simulation Models Because many population processes are stochastic, a PVA can never specify what will happen to a population. Rather, PVA can provide estimates of probability distributions describing possible fates of a population. The fate of a given population may happen to fall at the extreme tail of such a distribution even if the processes and probabilities are assessed precisely. Therefore, it will often be impossible to test empirically the accuracy of PYA results by monitoring of one or a few threatened populations of interest. Presumably, if a population followed a course that was well outside of the range of possibilities predicted by a model, that model could be rejected as inadequate. Often, however, the range of plausible fates generated by PVA is quite broad. Simulation programs can be checked for internal consistency. For example, in the absence of inbreeding depression and other confounding effects, does the simulation model predict an average long-term growth rate similar to that determined from a life-table calculation? Beyond this, some confidence in the accuracy of a simulation model can be obtained by comparing observed fluctuations in population numbers to those generated by the model, thereby comparing a data set consisting of tens to hundreds of data points to the results of the model. For example, from 1938 to 1991, the wild population of whooping cranes had grown at a mean exponential rate, r, of 0·040, with annual fluctuations in the growth rate, SD (r), of 0·141 (Mirande et a!. 1993). Life-table analysis predicted an r of 0·052. Simulations using VORTEX predicted an r of 0·046 into the future, with a SD (r) of 0·081. The lower growth rate projected by the stochastic model reflects the effects of inbreeding and perhaps imbalanced sex ratios among breeders in the simulation, factors that are not considered in deterministic life-table calculations. Moreover, life-table analyses use mean birth and death rates to calculate a single estimate of the population growth rate. When birth and death rates are fluctuating, it is more appropriate to average the population growth rates calculated separately from birth and death rates for each year. This mean growth rate would be lower than the growth rate estimated from mean life-table values. When the simulation model was started with the 18 cranes present in 1938, it projected a population size in 1991 (N ± SD = 151 ± 123) almost exactly the same as that observed (N= 146). The large variation in population size across simulations, however, indicates that very different fates (including extinction) were almost equally likely. The model slightly underestimated the annual fluctuations in population growth [model SD (r) =0·112 v. actual SD (r)=0-141]. This may reflect a lack of full incorporation of all aspects of stochasticity into the model, or it may simply reflect the sampling error inherent in stochastic phenomena. Because the data input to the model necessarily derive from analysis of past trends, such retrospective analysis should be viewed as a check of consistency, not as proof that the model correctly describes current population dynamics. Providing another confir- VORTEX Technical Reference 105 mation of consistency, both deterministic calculations and the simulation model project an over-wintering population of whooping cranes consisting of 12% juveniles (less than 1 year of age), while the observed frequency of juveniles at the wintering grounds in Texas has averaged 13%. Convincing evidence of the accuracy, precision and usefulness of PYA simulation models would require comparison of model predictions to the distribution of fates of many replicate populations. Such a test probably cannot be conducted on any endangered species, but could and should be examined in experimental non-endangered populations. Once simulation models are determined to be sufficiently descriptive of population processes, they can guide management of threatened and endangered species (see above and Lindenmayer et at. 1993). The use of PVA modelling as a tool in an adaptive management framework (Clark et at. 1990) can lead to increasingly effective species recovery efforts as better data and better models allow more thorough analyses. Directions for Future Development of PVA Models The PVA simulation programs presently available model life histories as a series of discrete (seasonal) events, yet many species breed and die throughout much of the year. Continuous-time models would be more realistic and could be developed by simulating the time between life-history events as a random variable. Whether continuous-time models would significantly improve the precision of population viability estimates is unknown. Even more realistic models might treat some life-history events (e.g., gestation, lactation) as stages of specified duration, rather than as instantaneous events. Most PV A simulation programs were designed to model long-lived, low fecundity (K-selected) species such as mammals, birds and reptiles. Relatively little work has been devoted to developing models for short-lived, high-fecundity (r-selected) species such as many amphibians and insects. Yet, the viability of populations of r-selected species may be highly affected by stochastic phenomena, and r-selected species may have much greater minimum viable populations than do most K-selected species. Assuring viability of K-selected species in a community may also afford adequate protection for r-selected species, however, because of the often greater habitat-area requirements of large vertebrates. Populations of r-selected species are probably less affected by intrinsic demographic stochasticity because large numbers of progeny will minimise random fluctuations, but they are more affected by environmental variations across space and time. PVA models designed for r-selected species would probably model fecundity as a continuous distribution, rather than as a completely specified discrete distribution of litter or clutch sizes; they might be based on life-history stages rather than time-increment ages; and they would require more detailed and accurate description of environmental fluctuations than might be required for modelling K-selected species. The range of PVA computer simulation models .becoming available is important because the different assumptions of the models provide capabilities for modelling diverse life histories. Because PVA models always simplify the life history of a species, and because the assumptions of no model are likely to match exactly our best understanding of the dynamics of a population of interest, it will often be valuable to conduct PVA modelling with several simulation programs and to compare the results. Moreover, no computer program can be guaranteed to be free of errors. There is a need for researchers to compare results from different PVA models when applied to the same analysis, to determine how the different assumptions affect conclusions and to cross-validate algorithms and computer code. Acknowledgments James W. Grier made available his simulation program, SPGPC, which provided many of the algorithms on which the first version of VORTEX (SIMPOP) was based. I thank Ulysses S. Seal, Thomas J. Foose, Jon Ballou, Nathan R. Flesness, Tim W. Clark, Gary Backhouse, 106 VORTEX Technical Reference David Lindenmayer, Simon Bennett, and many of the staff of the Department of Conser vation and Environment, Victoria, Australia, for many helpful comments on VORTEX and PVA. Tim W. Clark, David Lindenmayer and two anonymous reviewers provided valuable critiques of drafts of this paper. References Ak<;akaya, H. R., and Ferson, S. (1990). 'RAMAS/Space User Manual. Spatially Structured Population Models for Conservation Biology.' (Applied Biomathematics: Setauket, New York.) Ballou, J. (1983). Calculating inbreeding coefficients from pedigrees. In 'Genetics and Conservation: A Reference for Managing Wild Animal and Plant Populations'. (Eds C. M. Schonewald-Cox, S. M. Chambers, B. MacBryde and W. L. Thomas.) pp. 509-20. (Benjamin/Cummings: Menlo Park, California.) Belovsky, G. E. (1987). Extinction models and mammalian persistence. In 'Viable Populations for Conservation'. (Ed. M. E. Soule.) pp. 35-57. (Cambridge University Press: Cambridge.) Berger, J. (1990). Persistence of different-sized populations: an empirical assessment of rapid extinctions in bighorn sheep. Conservation Biology 4, 91-8. Brewer, B. A., Lacy, R. C., Foster, M. L., and Alaks, G. (1990). Inbreeding depression in insular and central populations of Peromyscus mice. Journal of Heredity 81, 257-66. Burgmann, M. A., and Gerard, V. A. (1990). A stage-structured, stochastic model for giant kelp Macrocystis pyrijera. Marine Biology 105, 15-23. Clark, T. W. (1989). 'Conservation Biology of the Black-footed Ferret. Special Scientific Report.' (Wildlife Preservation Trust International: Philadelphia.) Clark, T. W., and Seebeck, J. H. (Eds) (1990). 'Management and Conservation of Small Populations.' (Chicago Zoological Society: Brookfield, Illinois.) Clark, T. W., Warneke, R. M., and George, G. G. (1990). Management and conservation of small populations. In 'Management and Conservation of Small Populations'. (Eds T. W. Clark and J. H. Seebeck.) pp. 1-18. (Chicago Zoological Society: Brookfield, Illinois.) Clark, T. W., Backhouse, G. N., and Lacy, R. C. (1991). Report of a workshop on population viability assessment as a tool for threatened species management and conservation. Australian Zoologist 27, 28-35. Crow, J. F., and Kimura, M. (!970). 'Introduction to Population Genetics Theory.' (Harper and Row: New York.) Doughty, R. W. (1989). 'Return of the Whooping Crane.' (University of Texas Press: Austin.) Falconer, D. S. (1981). 'Introduction to Quantitative Genetics.' 2nd Edn. (Longman: New York.) Ferson, S. (1990). 'RAMAS/Stage. Generalized Stage-based Modeling for Population Dynamics.' (Applied Biomathematics: Setauket, New York.) Ferson, S., and Ak<;akaya, H. R. (1989). 'RAMASI Age User Manual. Modeling Fluctuations in Age structured Populations.' (Applied Biomathematics: Setauket, New York.) Fisher, R. A. (1958). 'The Genetical Theory of Natural Selection.' 2nd Edn. (Dover: New York.) Foose, T. J., Lacy, R. C., Brett, R., and Seal, U. S. (1992). Kenya black rhinoceros population and habitat viability assessment. (Captive Breeding Specialist Group, SSC, IUCN: Apple Valley, Minnesota.) Gilpin, M. E. (1987). Spatial structure and population vulnerability. In 'Viable Populations for Conservation'. (Ed. M. E. Soule.) pp. 125-39. (Cambridge University Press: Cambridge.) Gilpin, M. E. (1989). Population viability analysis. Endanger~d Species Update 6, 15-18. Gilpin, M. E., and Soule, M. E. (1986). Minimum viable populations: processes of species extinction. In 'Conservation Biology: the Science of Scarcity and Diversity'. (Ed. M. E. Soule.) pp. 19-34. (Sinauer: Sunderland, Massachusetts.) Goodman, D. (1987). The demography of chance extinction. In 'Viable Populations for Conservation'. (Ed. M. E. Soule.) pp. 11-34. (Cambridge University Press: Cambridge.) Grier, J. W. (1980a). A simulation model for small populations of animals. Creative Computing 6, 116-21. Grier, J. W. (1980b). Modeling approaches for bald eagle population dynamics. Wildlife Society Bulletin 8, 316-22. VORTEX Technical Reference 107 Harris, R. B., Metzger, L. H., and Bevins, C. D. (1986). 'GAPPS. Version 3.0.' (Montana Cooperative Research Unit, University of Montana: Missoula.) Hill, F. A. R. (1991). A research recovery plan for the brush-tailed rock wallaby Petrogale pencil/ala (Gray 1825). Report to Australian h!ational Parks and Wildlife Service. (Department of Con servation and Environment: Melbourne.) Kirkpatrick, S., and Stoll, E. (1981). A very fast shift-register sequence random number generator. Journal of Computational Physics 40, 517. Lacy, R. C. (1993). Impacts of inbreeding in natural and captive populations of vertebrates: implications for conservation. Perspectives in Biology and Medicine. (In press.) Lacy, R. C., and Clark, T. W. (1990). Population viability assessment of eastern barred bandicoot. In 'The Management and Conservation of Small Populations'. (Eds T. W. Clark and J. H. Seebeck.) pp. 131-46. (Chicago Zoological Society: Brookfield, Illinois.) Lacy, R. C., Flesness, N. R., and Seal, U. S. (1989). 'Puerto Rican Parrot Population Viability Analysis.' (Captive Breeding Specialist Group, SSC, IUCN: Apple Valley, Minnesota.) Lacy, R. C., Petrie, A. M., and Warneke, M. (1993). Inbreeding and outbreeding depression in captive populations of wild species. In 'The Natural History of Inbreeding and Outbreeding'. (Ed. N. W. Thornhill.) (University of Chicago Press: Chicago.) (In press.) Lande, R. (1988). Demographic models of the northern spotted owl (Strix occidentalis caurina). Oecologia 75, 601-7. Latour, A. (1986). Polar normal distribution. Byte August 1986, 131-2. Lindenmayer, D. B., Clark, T. W., Lacy, R. C., and Thomas, V. C. (1993). Population viability analysis as a tool in wildlife management: a review with reference to Australia. Environmental Management. (In press.) Mace, G. M., and Lande, R. (1991). Assessing extinction threats: toward a reevaluation of IUCN threatened species categories. Conservation Biology 5, 148-57. Maguire, L. A., Lacy, R. C., Begg, R. J., and Clark, T. W. (1990). An analysis of alternative strategies for recovering the eastern barred bandicoot. In 'The Management and Conservation of Small Populations'. (Eds T. W. Clark and J. H. Seebeck.) pp. 147-64. (Chicago Zoological Society: Brookfield, Illinois.) Maier, W. L. (1991). A fast pseudo random number generator. Dr. Dobb's Journal May 1991, 152-7. Mirande, C., Lacy, R. C., and Seal, U. S. (1993). Whooping crane conservation viability assessment workshop. (Captive Breeding Specialist Group, SSC, IUCN: Apple Valley, Minnesota.) Morton, N. E., Crow, J. F., and Muller, H. J. (1956). An estimate of the mutational damage in man from data on consanguineous marriages. Proceedings of the National Academy of Sciences, U.S.A. 42, 855-63. O'Brien, S. J ., and Evermann, J. F. (1988). Interactive influence of infectious diseases and genetic diversity in natural populations. Trends in Ecology and Evolution 3, 254-9. Possingham, H., Davies, I., and Noble, I. R. (1991). 'An Evaluation of Population Viability Analysis for Assessing the Risk of Extinction.' (Resource Assessment Commission: Canberra.) Possingham, H. P., Davies, I., Noble, I. R., and Norton, T. W. (1992). A metapopulation simulation model for assessing the likelihood of plant and animal extinctions. Mathematics and Computers in Simulation 33, 367-72. Ralls, K., Ballou, J. D., and Templeton, A. R. (1988). Estimates of lethal equivalents and the cost of inbreeding in mammals. Conservation Biology 2, 185-93. Resource Assessment Commission (1991). Forest and timber inquiry. Draft report. Vol. 2. July 1991. (Australian Government Publishing Service: Canberra.) Ricklefs, R. E. (1979). 'Ecology.' 2nd Edn. (Chiron: New York.) Robertson, A. (1960). A theory of limits in artificial selection. Proceedings of the Royal Society of London 153B, 234-49. Seal, U. S., and Foose, T. J. (1989). Javan rhinoceros population viability analysis and recommen dations. (Captive Breeding Specialist Group, SSC, IUCN: Apple Valley, Minnesota.) Seal, U. S., and Lacy, R. C. (1989). Florida panther population viability analysis. (Captive Breeding Specialist Group, SSC, IUCN: Apple Valley, Minnesota.) Seal, U. S., Ballou, J. D., and Padua, C. V. (1990). Leontopithecus population viability analysis workshop report. (Captive Breeding Specialist Group, SSC, IUCN: Apple Valley, Minnesota.) Seal, U.S., Lacy, R. C., Medley, K., Seal, R., and Foose, T. J. (1991). Tana River Primate Reserve Conservation Assessment Workshop. (Captive Breeding Specialist Group, SSC, IUCN: Apple Valley, Minnesota.) 108 VORTEX Technical Reference Selander, R. K. (1983). Evolutionary consequences of inbreeding. In 'Genetics and Conservation: A Reference for Managing Wild Animal and Plant Populations'. (Eds C. M. Schonewald-Cox, S. M. Chambers, B. MacBryde and W. L. Thomas.) pp. 201-15. (Benjamin/Cummings: Menlo Park, California.) Shaffer, M. L. (1981). Minimum population sizes for species conservation. BioScience 31, 131-4. Shaffer, M. L. (1987). Minimum viable populations: coping with uncertainty. In 'Viable Populations for Conservation'. (Ed. M. E. Soule.) pp. 69-86. (Cambridge University Press: Cambridge.) Shaffer, M. L. (1990). Population viability analysis. Conservation Biology 4, 39-40. Simberloff, D. A. (1986). The proximate causes of extinction. In 'Patterns and Processes in the History of Life'. (Eds D. M. Raup and D. Jablonski.) pp. 259-76. (Springer-Verlag: Berlin.) Simberloff, D. A. (1988). The contribution of population and community biology to conservation science. Annual Review of Ecology and Systematics 19, 473-511. Soule, M. E. (Ed.) (1987). 'Viable Populations for Conservation.' (Cambridge University Press: Cambridge.) Thomas, C. D. (1990). What do real population dynamics tell us about minimum population sizes? Conservation Biology 4, 324-7. Wright, S. (1977). 'Evolution and the Genetics of Populations. Vol. 3. Experimental Results and Evolutionary Deductions.' (University of Chicago Press: Chicago.) Appendix. Sample Output from VORTEX Explanatory comments are added in italics VORTEX-simulation of genetic and demographic stochasticity TEST Simulation label and output file name Fri Dec 20 09:21:18 1991 2 population(s) simulated for 100 years, 100 runs VORTEX first lists the input parameters used in the simulation: HETEROSIS model of inbreeding depression with 3 ·14 lethal equivalents per diploid genome Migration matrix: 2 0·9900 0·0100 i.e. 1% probability of migration from 2 0·0100 0·9900 Population 1 to 2, and from Population 2 to 1 First age of reproduction for females: 2 for males: 2 Age of senescence (death): 10 Sex ratio at birth (proportion males): 0· 5000 Population l: Polygynous mating; 50·00 per cent of adult males in the breeding pool. Reproduction is assumed to be density independent. 50·00 (EV=12·50 SD) per cent of adult females produce litters of size 0 25 ·00 per cent of adult females produce litters of size 1 25 · 00 per cent of adult females produce litters of size 2 EV is environmental variation 50·00 (EV =20·41 SD) per cent mortality of females between ages 0 and 10 · 00 (EV = 3 · 00 SD) per cent mortality of females between ages I and 2 10·00 (EV=3·00 SD) per cent annual mortality of adult females (2<=age<=10) 50·00 (EV=20·41 SD) per cent mortality of males between ages 0 and l 10 · 00 (EV = 3 · 00 SD) per cent mortality of males between ages J and 2 10·00 (EV=3·00 SD) per cent annual mortality of adult males (2<=age<=l0) VORTEX Technical Reference 109 EVs have been adjusted to closest values possible for binomial distribution. EV in reproduction and mortality will be correlated. Frequency of type I catastrophes: I ·000 per cent with 0 · 500 multiplicative effect on reproduction and 0·750 multiplicative effect on survival Frequency of type 2 catastrophes: I · 000 per cent with 0 · 500 multiplicative effect on reproduction and 0·750 multiplicative effect on survival Initial size of Population I: (set to reflect stable age distribution) Age 2 3 4 5 6 7 8 9 10 Total 0 0 0 0 0 5 Males 0 0 0 0 0 5 Females Carrying capacity= 50 (EV = 0· 00 SD) with a 10 · 000 per cent decrease for 5 years. Animals harvested from population 1, year I to year 10 at 2 year intervals: I females 1 years old I female adults (2<=age<= 10) 1 males I years old I male adults (2<= age<= 10) Animals added to population I, year I 0 through year 50 at 4 year intervals: I females I years old I females 2 years old I males I years old I males 2 years old Input values are summarised above, results follow. VORTEX now reports life-table calculations of expected population growth rate. Deterministic population growth rate (based on females, with assumptions of no limitation of mates and no inbreeding depression): r= 0·001 lambda=0·999 R0=0·997 Generation time for: females= 5 · 28 males= 5 · 28 Note that the deterministic life-table calculations project approximately zero population growth for this population. Stable age distribution: Age class females males 0 0·119 0·119 I 0·059 0·059 2 0·053 0·053 3 0·048 0·048 4 0·043 0·043 5 0·038 0·038 6 0·034 0·034 7 0·031 0·031 8 0·028 0·028 9 0·025 0·025 10 0·022 0·022 Ratio of adult ( > = 2) males to adult ( > = 2) females: I ·000 Population 2: Input parameters for Population 2 were identical to those for Population I. Output would repeat this information from above. Simulation results follow. Population! 11 0 VORTEX Technical Reference Year 10 N[Extinct] = 0, P[E] =0·000 N[Surviving] = 100, P[S] = 1·000 Population size= 4·36 (0·10 SE, 1·01 SD) Expected heterozygosity= 0·880 (0·001 SE, 0·012 SD) Observed heterozygosity= 1·000 (0·000 SE, 0·000 SD) Number of extant alleles= 8·57 (0·15 SE, 1·50 SD) Population summaries given, as requested by user, at 10-year intervals. Year 100 N[Extinct] = 86, P[E] =0·860 N[Surviving] = 14, P[S] =0·140 Population size= 8·14 (i ·27 SE, 4·74 SD) Expected heterozygosity= 0 · 577 (0 · 035 SE, 0 · 130 SD) Observed heterozygosity= 0·753 (0·071 SE, 0·266 SD) Number of extant alleles= 3·14 (0·35 SE, 1·29 SD) In 100 simulations of 100 years of Population!: 86 went extinct and 14 survived. This gives a probability of extinction of 0· 8600 (0·0347 SE), or a probability of success of 0·1400 (0·0347 SE). 99 simulations went extinct at least once. Median time to first extinction was 5 years. Of those going extinct, mean time to first extinction was 7·84 years (1·36 SE, 13·52 SD). 123 recolonisations occurred. Mean time to recolonisation was 4·22 years (0·23 SE, 2·55 SD). I 10 re-extinctions occurred. Mean time to re-extinction was 54·05 years (2·81 SE, 29·52 SD). Mean final population for successful cases was 8· 14 (1·27 SE, 4·74 SD) Age I Adults Total 0·14 3·86 4·00 Males 0·36 3·79 4·14 Females During years of harvest and/or supplementation mean growth rate (r) was 0·0889 (0·0121 SE, 0·4352 SD) Without harvest/supplementation, prior to carrying capacity truncation, mean growth rate (r) was -0·0267 (0·0026 SE, 0·2130 SD) Population growth in the simulation (r= - 0· 0267) was depressed relative to the projected growth rate calculated from the life table (r= -0·001) because of inbreeding depression and occasim;:zllack of available mates. Note: 497 of 1000 harvests of males and 530 of 1000 harvests of females could not be completed because of insufficient animals. Final expected heterozygosity was 0· 5768 (0·0349 SE, 0·1305 SD) Final observed heterozygosity was 0·7529 (0·0712 SE, 0·2664 SD) Final number of alleles was 3·14 (0·35SE, 1·29SD) Population2 Similar results for Population 2, omitted from this Appendix, would follow. ******** Metapopulation Summary ******** Year 10 N[Extinct) = 0, P[E] =0·000 N[Surviving] = 100, P[S] = 1·000 Population size= 8·65 (0·16 SE, 1·59 SD) Expected heterozygosity= 0·939 (0·000 SE, 0·004 SD) Observed heterozygosity= 1·000 (0·000 SE, 0·000 SD) Number of extant alleles= 16·92 (0·20 SE, 1·96 SD) VORTEX Technical Reference 111 Metapopulation summaries are given at 10-year intervals. Year 100 N[Extinct] = 79, P[E] =0·790 N[Surviving] = 21, P[S] =0·210 Population size= 10·38 (1·37 SE, 6·28 SD) Expected heterozygosity= 0·600 (0·025 SE, 0· I 15 SD) Observed heterozygosity= 0·701 (0·050 SE, 0·229 SD) Number of extant aiieles = 3·57 (0·30 SE, 1·36 SD) In 100 simulations of 100 years of Metapopulation: 79 went extinct and 21 survived. This gives a probability of extinction of 0·7900 (0·0407 SE), or a probability of success of 0·2100 (0·0407 SE). 97 simulations went extinct at least once. Median time to first extinction was 7 years. Of those going extinct, mean time to first extinction was I I ·40 years (2·05 SE, 20·23 SD). 91 recolonisations occurred. Mean time to recolonisation was 3 ·75 years (0·15 SE, 1·45 SD). 73 re-extinctions occurred. Mean time to re-extinction was 76·15 years (1·06 SE, 9·05 SD). Mean final population for successful cases was 10·38 (1·37 SE, 6·28 SD) Age I Adults Total 0·48 4·71 5·19 Males 0·48 4·71 5·19 Females During years of harvest and/or supplementation mean growth rate (r) was 0·0545 (0·0128 SE, 0·4711 SD) Without harvest/supplementation, prior to carrying capacity truncation, mean growth rate (r) was -0·0314 (0·0021 SE, 0·1743 SD) Final expected heterozygosity was 0· 5997 (0·0251 SE, 0·1151 SD) Final observed heterozygosity was 0·7009 (0·0499 SE, 0·2288 SD) Final number of alleles was 3·57 (0·30 SE, 1·36 SD) Manuscript received 4 March 1992; revised and accepted 13 August 1992